THE BOOK--Playing The Percentages In Baseball

I enjoyed reading all the articles, to some extent or other.

***

I think the idea behind the contest is extremely lame. 

At the same time, this lame idea brought a ton of talent out of the woodwork (one of the few times that the end justifies the means). 

What’s wrong with hiring all 10 of them, right now?  BPro pays by the article, not by year.  They are free to pay for as many articles as they want. 

***

Finally, I would immediately disqualify that article that appeared on MLB.com.  That’s more than ridiculous.  Why not let Matt and Brian link to all their existing great work each week, then?  This is like having an AI singer record their song in a pro studio, and submit that for their competition.

I agree on the MLB.com article.  I can’t understand how that was allowed.  It wasn’t even current, and didn’t give an impression that she’d do anything different from what “On the Beat” already does.

The winner will get the $1000 check and the crown, but BP has already been sufficiently impressed by the finalists that there’s a possibility that more than just the winner may be given an opportunity to write. From now on, the judges will pick the weekly topic, and we have to show a consistency in our quality.

As I said, there’s no reason NOT to sign more. 

“Sorry, Clay Aiken… you finished second… c u later.”

That’s not how it works.  The top 3-5 are signed to record deals, and I see it no different here.  The differentiator is the 1000$ bonus, which is fair, given the number of free articles BPro would otherwise be getting.

So, if Brian wants to work for BPro, he will.  It’s as simple as that.  I remember Aaron Gleeman once asked us about working at BPro as an intern.  I told him he was such a fantastically talented writer, that they should hire him on the spot.  I don’t know if he ever applied there.

This is like a hazing ritual.  Everyone gets in, but they make you stress over it first.  It’s somewhat (or is it absolutely?) sadistic on the part of BPro.

Put them out of their misery Will… when it gets down to the final 5, hire them all, and announce it.  Split the 1000$ among all 5.

Why did that guy who wrote about the Ibanez contract make the top ten?  His post was neither original nor very well written.  It seemed like they were just picking him by his bio, as the credentials he listed was longer than his actual article.  I’m not trying to sound petty, but it certainly takes the fun out of the contest if they are just going to pick guys who have a track record instead of guys who actually take the time to present a well researched and original article (not that the rest of the finalists didn’t do that). 

BTW.  I enjoyed your article a lot Brian.  I hope that either you or MattS wins, as you too seem ot be the most sabermetrically oriented, and BP could use a couple more guy like that.

Didn’t Colin turn in an article?  It’s hard for me to believe that his wasn’t among the best.

Right, I read that Colin did write one.  Just proves my point about the amount of quality out there.  “Sorry Colin… we really need quality analysts, but we can only hire one.”

***

Will Carroll comments on Brian’s article:

BP staff member Will Carroll
BP staff
(1112)

I’m adding my judging comment to each article:

Cartwright, Brian—7. I think this is a BP quality piece. Most of it is way above my head and therein lies my problem with it. I just don’t grasp most of it and he makes little effort to bring it down to a level where I get any sort of takeaway or even an explanation. As an example he starts off with a question about “elapsed time” - I’m not sure what it means and I don’t see where he answered it in the piece. There’s going to be an element of BP readers that loves this and I don’t blame them, but there’s going to be a lot of people that just don’t get it.

Clearly, Brian’s article was a niche piece.  So, it must be judged on that. 

And “BPro” quality piece?  What does that mean?  I’ll say it’s an MGL-quality piece.  This is standard MGL stuff, which is far far higher than the standard BPro stuff (insofar as research, analysis, and insight goes).

I sent this to Will as well:

Regarding the elapsed time:

Imagine a bell-curve of talent level for a player from age 20 to age 40.  A player improves, peaks, slows down, and then disintegrates.  You pick a spot at age 23 and a spot at age 33, and he likely won’t be the same kind of player.  The commonality in this case is simply his name.

But, if you have someone at age 23 and age 24, then we can figure that his talent level has not changed much.  The less “elapsed time” between performance, the more static is the player’s underlying true talent level.

This is why when we try to do these MLEs, we try to link a player’s minor league stats to his most recent (or same-season) MLB stats.

All congrats to Brian, Matt, and the others.

I admit—Colin not making the finals is my fault. I told several people (including Colin, I think) that he was a lock before it was announced, based on my awesome talent-evaluation skillz.

So,yeah, I jinxed it.

Seriously, though, I really can’t believe Colin didn’t make it. As far as quality in general, I’d have to say that (for example) his pieces on run estimators at StatsSpeak and THT, BsR-FIP, “The Root,” VORP (hmmm...) etc., along stuff by him and a lot of other people (Brian’s stuff on park factors, Matt’s on BABIP) is, in my (valueless) opinion, better than anything I’ve read from BP in a long time.

This isn’t to take anything away from the people who did make it, though. Haven’t read all of them yet.

BP is now the fourth different site where I have been published. What I find most difficult in sharing my thoughts is determining the level of comprehension of each audience. Sometimes you throw out a phrase assuming people will know wht you mean. On the other hand, the article can bog down if you spend too much time explaining every detail (I did have a couple proofreaders, a guy in the office and an online friend).

I said in my intro that I miss Dan Fox’s writing, and I would like an opportunity to attempt to fill that void. Not everyone reads those kinds of articles, but many of us do. 

Being assigned topics (this weeks is fairly general) will show how much breadth of knowledge we have, as well as being able to write well on demand.

and a couple things I remember from appearing in a real court “Don’t confuse the judge - don’t annoy the judge”

Good luck to both Brian and Matt—nice work in both cases.  (The third piece I read was “Back of the Envelope” on why TTO players have low BAs.  It’s very weak.)

Leaving aside the merits of Will’s comments, my suggestion to Brian was going to be to try to simplify his presentation a bit in the weeks ahead (and this was even before I read the BPro reader comments).  You are deeply immersed in sabermetric methods and lingo, but if you want to get votes in this competition you’ll have to reach an audience that isn’t. 

A few thoughts (for Matt as well):
*Be sure to tell readers in the very first paragraph why they should care about what follows—in this case, better player projections (that help them win their fantasy league!). 
*Keep tables simple.  They are there to illustrate a point, but don’t need to be comprehensive. (You don’t need to show doubles, triples, and doubles-triples combined.)
*Have a non-SABR baseball fan review your pieces and give you feedback from layman’s perspective before you submit them.

I desperately hope that someone more statistically inclined gets picked.  BPro holds no interest to me with the departure of Fox and other true stat-oriented writers.  The fact that someone like the MLB.com writer made it though goes to show how far the staff at BPro has drifted from stats to just generalized reporting.

I was really disappointed when Dan Fox left (SFR was one of my favorite recent reads at BPro) and I’m probably not renewing my subscription on the next cycle. 

I also found it mildly ironic how the stats-oriented pieces were using data from sites like fangraphs or pitchf/x.  Again, it just goes to show how far they’ve drifted toward irrelevancy.

I feel much the same way as azruavatar/#12.  So far this year, between Clay Davenport and Nate Silver, there have been 9 total articles.  Woolner’s gone and Fox is gone, and they haven’t been replaced (especially Dan Fox.)

I don’t feel like I’m really in their target audience anymore, so I didn’t renew my subscription the last time around.  There was a time when they had a lot to teach me, but I feel like that time has come and gone.

A smattering of comments…

azruavatar/#12 - I don’t think the lack of BPro stats in the entries is a result of their irrelevancy, but more a byproduct of our inability to replicate many of their most noted metrics because of the “black box” they tend to work behind.

I’d like to see some of the other judges, not just Will, post their grades too.

Outside of the local favorites here (Brian and MattS), I thought Tim Kniker’s article was rather good, though I’d like to see that analysis over a longer period of time than a single season.

Brian, was your article the same one you put up on Seamheads? Your archive there has seemed to disappear, and I swear you put that one up while you wrote there.

I have read Brian’s and Hissey’s so far.  I am a big fan of Brian’s work in general, but I don’t think this is one of his better pieces for several reasons.  My biggest criticism is that it is written WAY too far over the head of the average BP reader.  Even a technical article needs to be written in “the lowest common denominator,” and I don’t think this was.  In other words, even given the rigor of the material, I don’t think it was written clearly enough.  I don’t think that more than 5 or 10% of BP readers would get much out of this article.

And, as someone pointed out in the comments section, one of the first rules of writing is to let the reader know what you are going to do at the beginning of the piece.  I don’t think Brian did that.

Hissey’s piece was OK, but weak.  It was not written particularly well, certainly not up to professional standards, but the content and approach were more appropriate for the average BP reader I think.

Tango mentioned that BPro pays by the article.  I’m wondering if anyone here knows what the going rate is for an article on BPro (or any other paying site for that matter).

I seem to remember that about 3-5 years ago, 40$ was their standard.  And now it’s at least double that.  I wouldn’t be surprised if there’s at least a few writers at over 100$ an article.

That’s a decent deal, but seeing that turnover is rather large, and their subscription base is rather large (probably takes in half-a-million bucks), it doesn’t seem equitable.  But, it seems to work.

In contrast, Fangraphs started paying for bloggers right away, and he’s only got Google Ads.

The Funck piece addresses an interesting topic:  a possible link between hitters’ facial dimensions (width:height ratio) and hitting performance.  But it falls far short of presenting a compelling case:
* The writer estimates the ratios after already knowing which hitters are the most/least patient, introducing possible bias (he should have had someone else do the facial measurements)
* We aren’t told if this facial metric correlates with other things that might explain differences in hitting ability (e.g. height, weight). 
* Even if there IS a relationship, we don’t know if it’s really patience that’s the key factor (as opposed to power, or something else).

Still, it’s an interesting avenue for study....

I have no idea how they picked the MLB.com piece as a finalist, and not whatever Colin wrote. Any of his articles for THT should have qualified, and most of them for StatSpeak would have fit. There was a piece on BtB that didn’t make it that I thought was better than at least 3 of these articles. The Ibanez article wasn’t any better than the FanGraphs blog posts on the same subject (then again, some of those were written by Eric Seidman).

I didn’t read all of them (I read more than half), but I really expected better. The “payroll by position” article I thought was pretty pointless.

Read the Knight one on TTO players and BA.  I read all the comments and I am surprised that no one stated the obvious:

Are we supposed to be surprised that a player that makes a lot of outs is never the batting champion?  Isn’t that essentially what he is looking at?  It has nothing to do with the fact that these are K.  In fact, a player with a lot of K will have a higher BA when he does not K than a player with fewer K, which is sort of the opposite of his thesis.

This article is really silly.  The fact that Will really likes it leads me to be quite skeptical as to the basis of BP’s judgment in selecting the winner.

BTW, I think this is a good idea from a marketing perspective, which is the only thing that counts for BP.  Look at the interest it is receiving on this blog alone…

I did submit a piece; right now the plan is to do a bit of editing and use it as my THT piece for this week. Thanks for the kind words, everyone.

I was initially against this idea somewhat, because it came across a bit arrogant.  “Hey, we’re too lazy to go recruit talented writers, so come to us.” I’ve backed off that somewhat, plus I’m really enjoying all the extra articles right now.

For the record, if anyone’s interested in writing for Beyond the Box Score, you can just email me any time and ask “Hey, are you guys interested in having me write for BtB?” And, because of the cool way the SBN software is set up, anyone can publish their own work at any time in a FanPost, which will get nearly as much attention from the BtB community in the comments as a regular post, and if it’s good, it’ll get bumped to the front page anyway and be seen by everyone at other sites.

I don’t have $1000 to give away or $75 per article to pay, however.

It doesn’t need to be said by me, Colin, and I didn’t read your work, but if BP can’t look at the work you’ve done and realize you’d immediately be one of their biggest assets then the site deserves the criticism it gets.

If Colin had sent in his PECOTA-is-wrong-about-Wieters piece from THT (which was definitely better than any of the finalists’ submissions), would BP have simply brushed it off as an insult to PECOTA and Davenport?

weskeleton/#16: BP mentioned in the introduction to the competition that the winner will be paid $75 per article. Surely, the rest of the staff gets paid more.

I definitely don’t think those were the best ten articles submitted.  It seems as though BP was trying to get a repristentative sample of different kind of writers.  They only had a couple of stat guys (Matt Swartz and Brian), they had a couple of guys who might be interesting (Jeremy Greenhouse from Baseball Analysts and Kniker), but the rest of them seemed like MLB.com peices. 

The Ibanez was mediocre and unoriginal, yet Will rated it higher than Brian’s.  He mentioned in the comments that he like his personality.  I guess this means they were looking more for style than substance.  The “payroll by position” article was well written, but it wasn’t that interesting of a topic.  Will rated that one a 9.

To be honest, I expected the list to contain a few more names that I knew from around the Sabersphere.  I’m rather happy to see Matt and Brian in there (yay StatSpeak!), along with Jeremy G.  The Kniker article I thought was pretty good.  (In Lescroart’s bio, he took a rather cheap shot at THT and Fangraphs specifically, which was not deserved.) I had a few names in mind that I expected to see there, but didn’t.  My first throught was “Maybe they didn’t submit anything...”

I suppose that if we look at this like the draft, it would make sense that BP would be drafting statistically-minded folks, since they’ve lost a bunch of them (almost all of them?) over the past few years to “free agency.” For BP’s sake, I hope that one of the more stat-minded guys wins.  And if not, that they give them an offer.  But BP seems to be tilting toward being a more general site lately, and that’s their prerogative.  I suppose their list reflects that fact.

A telling quote from Will Carroll in the comments section:

“Kniker, Tim—8. This article reminds me of Keith Woolner, which is pretty damned high praise. He’s not as good a writer yet, but he’s certainly good and the process where he gets places is very solid. I worry a little bit that he’s all research and he’ll get squashed by some of the themes, but I can’t wait to see him try.”

While that’s not cut and dry, it kind of says that they’re not just looking for someone research-oriented to replace Woolner, Fox, et al.

dan/#28,

I think it just goes to show who your faces of BPro are these days: Carroll and Christina Kahrl. Perhaps not coincidentally, both are BBWAA members. Kevin Goldstein was the third judge. All 3 do more reporting-type work, so maybe the fact that BPro Idol isn’t stacked with our sabermetric friends isn’t such a surprise.

Matt/#14 - Their stats are either riddled with flaws, black box or the most basic of stats you can get anywhere else.  I check the minor league equivalencies from time to time but I don’t think I’ve done more than that in the past year.  Whichever of the three categories they fall into, that makes them irrelevant in my book.  As #13 said, I just don’t think that I’m their target audience anymore.

dan/#28 - I was a little perturbed by Will posting his comments.  It needed to be an all the judges or none of the judges thing, imo.  I’d rather it had been none of the judges to avoid tainting the voters for or against certain writers.  Also, Will came off as exceedingly arrogant in several of the comments—they were abrasive in tone and condescending in substance. 

The worst from the Ken Funck piece—“It’s a piece that Huckabay—the obvious influence—would have knocked out of the park. This guy? Double.” That maybe the way Will really feels but the complete lack of tact was distasteful.

I think it just goes to show who your faces of BPro are these days: Carroll and Christina Kahrl.

I believe I read that Joe Sheehan was not involved in the latest BP annual.  Maybe someone can confirm (or dispute) my recollection.  There always seems to be some in-house fighting over there.

***

azru/30:

Also, Will came off as exceedingly arrogant in several of the comments—they were abrasive in tone and condescending in substance.

The worst from the Ken Funck piece—“It’s a piece that Huckabay—the obvious influence—would have knocked out of the park. This guy? Double.” That maybe the way Will really feels but the complete lack of tact was distasteful.

This is exactly the difference between azru and Will.  azru gave his opinion, and provided the evidence to support him.  Will on the other hand gave a summary opinion, and therefore can only be read as being insulting.  That’s the difference between being insulting and offering constructive criticism.

Anyway, every time I see Will make a comment comparing anyone to BPro, I have to remember to skip over it, since his Paula Abdul cheerleading comes through.

***

In Lescroart’s bio, he took a rather cheap shot at THT and Fangraphs specifically, which was not deserved.

I skipped over all the bios.  Let’s see what Pizza is talking about…

I believe what separates BP from some of its (dare I say lesser?) rival sites such as THT or even FanGraphs is the quality of its writers. There are numerous sites and publications today with brilliant statistical minds making interesting inroads into the “science of baseball,” but as far as I know only BP has been able to combine forward-thinking, sabermetric genius with talented writers who not only make content accessible, but also completely enjoyable to read. Suffice to say, it is no coincidence that we are seeing mainstream sites like ESPN.com pick up BP’s work as opposed to other leading statistical sources due to the fact that only BP’s writers can stand up to the rigors of mainstream journalistic appeal.

Wow, what an a-hole comment to make.

And then his article is based on the great data from Fangraphs.  How in the world was he able to make any sense of the data from Fangraphs, if all the genius forward-looking writers are at BPro?

Again, people should learn how to speak without being insulting.  Provide EVIDENCE for your opinion.  That’s all you need to do. 

MGL for example goes off on lots of people, but at least he supports his position with evidence.  You can understand why he says what he says.  It’s ok to have an opinion, like Miss California.  It’s not ok to have an unsubstantiated opinion: what’s to stop you from changing said opinion on a whim?

***

And I would like to see the judges’ comments (all three of them) all on one page.

Wow, great to see the support from everyone here!  Thanks for the suggestions from everyone.  I was also a little surprised that there weren’t more people that I’d heard of, and I’m sure whatever Colin submitted should have been picked at least over mine and probably many others (unless he submitted his Wieters article, which obviously might not be good for business at BP!), but I still think the competition is fun and a symptom of BP reengaging the sabermetric community.  Many baseball fans have heard of BP, but they don’t really know about other sites since they aren’t that into it, and grabbing some people from other sites will help spread the word that there are many sites out there that they can check out. 

It also serves the much larger goal of making more people want to do sabermetrics.  Some of the finalists not regular bloggers, and I’m sure there were hundreds of entries from people taking their first stab at their own sabermetric research.  That not only means more intelligent people are out there looking for an outlet, but also means more intelligent people learning about what is done elsewhere on the web and improving what current bloggers are doing. 

From a personal standpoint, there’s something awesome about all of my family and friends suddenly wanting to learn about sabermetrics and read the stuff I write.  The sentiment went from “it’s great that you have a cool hobby, Matt” to “You’re a finalist?  Send me a link!”

Well, if I was supposed to be one of the “brilliant statistical minds” at FanGraphs who was not yet a talented writer, I can accept that. I always have struggled with the writing part of it, even in college. I had the same thoughts as Tango, that it came of as an insult to FanGraphs, then he constantly referenced their stats. Maybe BP has the writers, FanGraphs has the stats library.

Brian,

It seems to me that there’s a huge selection bias problem in your article, in that the MLEs are being tested on the same data set they were developed with. In that case, of course the direct MLEs will work better, regardless of their actual quality. What you need to do is test out of sample, and perhaps even using MLEs from one year and MLB stats from the next (if the difference between direct and chained MLEs is caused in part by selection biases that arise from why players get called up straight to the major leagues in the first place).

Does anyone else get the sense, reading some of his comments, that Will is not only arrogant but, um, perhaps somewhat lacking in the old smarts department?

Maybe BP has the writers, FanGraphs has the stats library.

You are being too nice.  Have we learnt nothing from racism, that you don’t paint everyone with the same brush, just because their share one common trait? 

***

As for testing MLEs, here’s one way to do it: come up with your forecasted WAR for every minor league player, through age 29.  Come up with the 100 best minor leaguers in any given year based on the forecasted WAR through age 29.

Do this in a competition with other systems (including scouts).

Who wins?

I would bet that the “pure stats” ones would not do so well.

This is almost identical to my Forecast Challenge 2009 (which reminds me, I need to put up a small update), except instead of just one year, you are looking at one to ten years of future performance.

Tom, I completely agree with the first comment here, in that at the very least this contest showed how many talented people are out there.  If one knew web design, heck, six to seven of them could start their own site and probably do well.

I got into some of the articles, and there were others that didn’t really hold my interest.  The mlb.com thing was a bit bothersome, and I’m unsure how the Ibanez article, of which I personally wrote 3-4 different variations of in Fangraphs-style posts, could beat out hundreds of other submissions.

And taking cheap shots at other sites does not sit well with me either, and I write for BOTH Fangraphs and BP, hoping both do well.  Someone can praise one site without insulting another.

Dave/34

There may be a bias, but I believe chaining has other biases that are much greater.

The sample of 368 rookies had a combined Marcel slash line of 270/327/427 for their first 2-3 years in MLB. MLB Direct, for High A or lower, projected them at 267/323/419. MLB Chained projects same players, using the same stats at 242/295/368. These two models were both calculated from virtually the same set of players who eventually played in MLB. The differences were not using chaining from A+ to AA to AAA to MLB, and not using a strict consecutive years rule for matched pairs (which necessitates using chaining for low minors). I would conclude that each additonal step of chaining multiplies the amount of the bias error at each level.

The sample of 368 rookies had a combined Marcel slash line of 270/327/427 for their first 2-3 years in MLB.

This is expected based on the definition of Marcel, which gives every player 240 PA of league average PA to “start” their careers with.

***

You can find a hole in every single MLE paper or article out there.  It’s simply something that you have to live with.  For those who’ve never seen it:

http://www.tangotiger.net/hateMLEs.html

***

By the way, this was one of the first articles I ever posted on my site.  I think it would have fit in with the general BPro theme:
http://www.tangotiger.net/SpeedLead.htm

I remember an unknown named Nate Silver posting on Primer if anyone ever did tools-based sim scores as a way to figure aging curves.  I sent him that link, and told him I’d be surprised if he could find anything.  Several months later, PECOTA came out.  I’m still skeptical.

In the chaining method, are you accounting for sample bias?

Say a guy hits .360 in AA, goes to AAA and hits .235.  He probably wasn’t a true .360 at the AA level.  The group of players who are selected to be called up are, as a group, lucky players.  You have to start by regressing their stats somewhat to the league mean.

Rally/40 - In the article I detail; how I ran three different models tht use chaining, to try to account for different types of biases. One mthod used every player who played in High A, compared to what they did in AA, AA to AA, AA to MLB, in consecutive seasons (year-1, year0, or year=1). The next had the same time constraint, but limited to players who had 2.5 PAs per game, to avoid usng bench-riders and stats heaving on pinch hitting. The third used 2.5 PA per game, but only players who eventually played in the majors.

In Oliver, I regress to the league mean as a final step. Normalize to park, normalize to league, combine into a single line for each year, rgress to highest level appeared in, apply age factors.

Tango/39 - For the BP article, I used Marcel in the year after their rookie season, using 3*rookie-1, 4*rookie, and 5*rookie+1. The players had to have 150+ PA in their rookie season. I thought this model would give the players a large enogh sample of their own stats while givng a good picture of their true talent when entering MLB.

Rally/40 again
Step 1 for true talebt - I have a db of all minor leage stats 1998-2009. The ‘A+’ number I was comparing to the rookie Marcel was a 5-4-3 weighted projections based on minor league stats in a season where a+ was the highest level achieved and 50% or more PA for the year were at that level. I calculated all levels, published A+ and higher.

Step 2 - I sed as many players as possble in the sample and control groups, using weighted matched pairs of what they did at each level, so even if I don’t catch each players true talent, I based the calculations on the group as a whole.

There were only 2 articles I thought were good, Brians and a guy named Tim Kniker, although maybe it was the topic for most of the others that biased me.

One writer claimed sluggers BABIP was higher because their HR could not be caught and another said pitchers have no control over BABIP, and we know some do. I never finished their articles so maybe I misunderstood in my quick screening.

It looks like they are using the contest to see what the voters who are subscribers like as well as to collect articles, even those who did not make the final 10.  And for those writing a team article, I am surprised they were finalists. I mean, anyone can write about their favorite team and that’s not what BP is known for. Maybe they need writers for the next BP annual. Hope they print it with a player index this time

I have know Tim for a while via a Royal’s message board.  He is a very smart guy, that whole PhD at MIT thing.  I am pretty sure he works pretty hard at his day job and it should interesting to see what he can put out if he concentrate on the writing.  I am pretty sure he hasn’t written for other pubs so he should have some pretty original ideas.

Tim’s article was indeed interesting.  However, I think he would have been better off defining his position “starters” (and top starting pitchers) based on their Opening Day status, rather than the season performance that followed (which is what I think he did).  By using PAs and IPs, he includes among his starters some players who may have been bench players, or even in the minors, when the season began.  So it’s not that surprising that he finds that starters on both playoff and contending teams make equal contributions (since hindsight allows us to know they got the most playing time).  I’d like to know how many of those players were starters when the season began.

He’s still looking at something interesting, which is how much value teams get from their top 9 vs. other players (or top few pitchers vs. other pitchers).  But if he wants to analyze “organizational depth” I think he should look at the contributions of players who weren’t expected to play a major role as of Opening Day, compared to those who were.

Right.

And next year, he could use this:
http://tangotiger.net/survey/

A note to Brian:  I could see some of the BPro readers/voters possibly being put off by your commenting on the articles of your competitors.  I don’t mean your intentions are in any way dishonorable, but I could see some people thinking it’s inappropriate since you are part of the competition. 

Anyone else have this reaction?

Yes. 

I think fair-minded folks like us appreciate it.  But, Brian is an unknown to most of those BPro readers, so he doesn’t have the backstory to do it.

This is a case of “appearance of conflict of interest”, even though we (the readers here) can agree that Brian would be able to disassociate the competitor from the commenter.

Brian, I’m not getting it.  Perhaps you could explain what you are doing in chaining by using a single player as the basis for calculation and showing an example?

Someone over at the Royals board mentioned this discussion thread, so I thought I’d come over and say hello and introduce myself.

Thanks for the nice comments.  I feel honored to be considered in the same ballpark as some of your hometown favorites (Brian & MattS).

To confirm, it’s true that I’ve never written for another pub/website, and in fact what I sent to BP was the first article on baseball that I’ve ever written.  Though I have done a lot of posting (sometimes the equivalent of a 500 word piece) over at the Royals message board and talked a lot about baseball modelling with a classmate of mine at MIT, Joel Sokol, who is now a professor at Ga Tech and sometimes comes here.

I have a number of ideas that I hope to try out in the competition, but it’ll have to depend on how well they fit within the given theme, and if I have time to do the rudimentary analysis needed, given that essentially each article needs to be researched and written in 3 days (typically we will get our “theme” Tuesday afternoon and have to hand in a piece by Friday noon).

If you have any feedback, I’m always glad to hear it.  There’s an ongoing thread at the Royals scout site, or I’ll likely check over on this thread every week or so.

Nice to meet you all

I’m glad to see Inside The Book is really getting excited about BPro Idol. It should be a great contest, and I can’t wait for a superstar to emerge after 10 weeks!

If Colin had sent in his PECOTA-is-wrong-about-Wieters piece from THT (which was definitely better than any of the finalists’ submissions), would BP have simply brushed it off as an insult to PECOTA and Davenport?

Don’t discount the fact that Colin’s PECOTA is wrong piece may have cost him any shot at placing in the contest.

At a book signing with Clay Davenport, Steve Goldman, Jay Jaffe and Ben Lindbergh, I stood up during the public Q&A and specifically challenged Clay on PECOTA’s projection of Wieters. He gave his best answer, did say the his own DTs were ‘quick and dirty’ compared to PECOTA, but had no hard feelings as I spent a long time afterwards talking to him about details of projections, park factors, etc.

Rally - I won’t be able to give you that detailed of a response until Friday or Saturday. After work tonight I have to do the first draft of next week’s BP piece, the final draft tomorrow night.

You might be right, sg - we may never know for sure. (There’s even a non-zero chance that Kahrl, Goldstein and Carroll have never read the article, although I sent Goldstein some of the research in it ahead of time and never got a reply.)

I think it’s far more likely that they simply were grading for very different things than, say, a judging panel of Rally, Tango and Guy would. A lot of Carroll’s comments focused on things like tone and personality rather than content, for instance. (Joe Sheehan, to his credit, did post this in the comments of one of the finalists: “Relentlessly mislabeling BA on contact as ‘BABIP’ is a massive hole.” Carroll, meanwhile, said “My first 9! Knight certainly has the stat chops and does the first heavy number article I see. He lacks a little bit of readability, but if you put in a couple scatter charts, this could be a Nate Silver-style article.") I think there’s plenty of room here to acknowlege a different point of view and then disagree with it, rather than trying to assign any other motive to it.

Well, I have to say that Colin’s current piece at THT (http://www.hardballtimes.com/main/article/who-watches-the-watchers/), just as one example, is better than most if not all of the “Idol” submissions.  So it does seem odd he didn’t make the cut.  But I guess BPro’s loss is Studes’ gain.

Yeah, but we had him first, so Studes’s gain is StatSpeak’s loss!

Yes, that was a fine article by Colin. Thanks for the sim score formula - I had Pythagorean scores, but hadn’t figured out how to convert them to the 100 to 0 scale yet.

Thanks for the kind words, gentlemen. That is in fact the article I submitted to BP Idol (or a revision of it, at least - at THT I have the benefit of an editor, which isn’t so with the Idol contestants).

PC, I’m not done with StatSpeak just yet, just in… hibernation. Eventually here I plan on doing some more SQL tutorials, which I don’t think works for the THT audience. (What I’d really like to do is publish the SQL for full Retroera WAR for hitters from scratch, with a tutorial, but that’s a tall order.)

And Brian, the formula for the sim scores is… junky, in my estimation. Which isn’t to say it doesn’t work, but I can’t say I’m fond of it. (I really dislike arbitrary constants - and there’s three of them!) I’d love to hear a better approach to the matter, if there is one.

I do somewhat of a similar “junky” sim score for my scouting report.  Works ok for what I need.

I should probably (and probably will) send this to the Pitch/FX gurus, and possibly Will himself, but it was my understanding that Josh Kalk’s web tool didn’t spider the data directly (a la Brooks’ ) but accessed Josh’s own data.  Which is why, if you look back at his blog, you’ll see he would occasionally note that the tool was “a few days behind” reality.  And judging by the addresses of his 08 tool and 07 tool, I’m almost certain that the Hafner article that purports to analyze Pronk’s revitalization this year, is actually using the 2007 Pitch/FX data. 

That’s not just sloppy, but it also makes the analysis—“In conclusion, it looks like at the very least we are looking at a healthy Travis Hafner for the first time in 2-3 years.”—almost laughably incorrect.  The data that he says shows a new Hafner is in fact the data from the old Hafner.

Again, I’m not 100% sure that’s how Josh’s tool worked, but I’m fairly certain.

Indeed, Josh just emailed me back and says that his web tool has no 2009 data. 

Kind of a big mistake.

Wow, yeah those pitch f/x graphs in the Lescroart article are definitely not from 2009 data.  He claims Hafner has seen 61 95+ mph pitches in 2009, but he has only 73 PAs.  Shouldn’t he or someone at BPro noticed this.  And if you look at the swinging strikes image, there are at least 60 dots there.  Again 60 swinging strikes in 73 PAs should set off alarm bells.  I have 27 swinging strikes and 11 pitches seen over 95mph in 09.

I know Tango mentioned it but it is so annoying that he calls FanGraphs and THT lesser sites in his bio, then in the articles uses all FanGraphs stats and graphs “by Josh Kalk for hardballtimes.com”

Dave- I’m sure he was just sucking up to B-Pro.  Not the most admiral strategy, but a smart one if you want the job.  It is rather hypocritical that he would use data from FanGraphs though. 

He wrong obviously.  FanGraphs is much better IMO for getting stats then B-Pro, as they have a much easier sortable database, and they have “better” stats.  And THT is clearly a better analysis site.  Guys like Turkenkopf and Wyers could write circles around most of the guys from B-Pro.

What really bugs me about the putdowns is that I always use Fangraphs stats in my BP articles.  When I was first hired, I asked Joe Sheehan if there were requirements to use BP stats and he said, verbatim:

“Eric, use whatever numbers you feel best tell the story.”

And I had similar responses from others, too, so the usage of FG or THT stats isn’t a bugaboo for them and yet this author purposely put the sites down… and since I write for Fangraphs and have worked really hard with Dave Cameron, David Appelman, Marc Hulet and the rest of the staff, I take it very personally.

To then go out and use stats from the site is absurd.

“Eric, use whatever numbers you feel best tell the story.”

I don’t know about others, but I think that says a lot about the direction of BPro. I wonder if they are even interested in sabermetric research to the extent most of us here would like them to be, or if they want to be the place that uses the newer metrics to tell baseball stories.

Matt, I think you’re reading that the wrong way.  “Tell your story” is essentially the shorthand for making the article presentable and well-written.  They were basically giving me free statistical reign to get my point across… if it meant using xFIP from THT instead of their QuikERA, so be it.

If you read some of the comments to my audition article, there are readers who don’t think that BP is the place for stat articles. I had to explain how ths wasn’t so not so long ago.

Eric/#67,

I’ll take your word on that. It is probably unfair for me to judge based on that isolated comment.

Nevertheless, I think the instinctual reaction is worth noting in this case. It seemed to me when I was first introduced to BPro a few years back (c. late 2005) that their writers would use nothing but stats that were either from prior research (James, Palmer, et al.) or were from their own writers (VORP, WARP, JAWS, etc.). It was as if they made a conscious decision to push their own research. Clearly, if that was true, then their stance has changed a lot.

My impressions could easily be wrong about this, so feel free to correct me if I have stated something in error.

@Brian

Carroll’s comment in that thread about doing it every day is one of the funniest things I’ve ever read.  Not only does he not get that it’s a quote from the article, but the suggestion that reporting tittle tattle and gossip from trainers is somehow the same is almost insultingly stupid.

Are there any legal issues with using FanGraphs or THT stats in an article for BP that you are getting paid to write (and BP subscribers are paying to read)? I know those stats are freely available for anyone to look at, but repackaging them for profit seems like it could be questionable.

Excellent work as usual, Brian.  Well done.

If I were still a subscriber to BP, I’d say the same over there and vote for you, but I’m not, so I can’t.  Don’t let the naysayers get you down.  I’ve had my share of naysaysers even for my very best articles.

This week’s articles are now up...let the voting begin

http://www.baseballprospectus.com:80/bpidol/

Some real quick initial comments; more to follow.

Kniker - Overall solid. I would have loved to see a bit more explanation about the potential for sampling error in an RE matrix, but I understand how that might have been beyond the scope of the article. Well presented and a good handle on the technical aspects.

Funck - Very solid, but again sparse on some of the finer details. (Again, my preference here seems to lean away from the judges’, and so shouldn’t be taken as necessarily a problem.) I think in the format presented it would do better to be clearer on if he’s focusing on EqA or on the Translations. But again, he seems to have a handle on how to do this sort of work.

Ghiroli - Maybe it’s written well - I’d have to really put aside my other reservations to try to figure out what I thought there. As a stat piece it’s rather lacking, and that’s putting it kindly. The article is filled with unsubstantiated statements, like “Because batting average is an unstable statistic from year-to-year, the rate at which a player hits ground balls-to-fly balls (G/F) has become an important barometer for teams evaluating new prospects and current talent.” The article seems to cherry-pick players and look at their G/F ratio to try to divine any meaning. She wraps up by saying “When used in conjunction with a player’s specific tendencies at the plate and power potential, G/F can be used to project offensive efficiency, alter mechanics, and protect against the statistically weakest batted ball, the infield fly,” which at no point is established in the article (and the way she presents it in the article, I don’t even think it’s true). It’s overused and abused in sabermetrics, but so much as a correlation coefficient here or there would have gone a long way to helping this article out (with the big caveat that I’m not sure it would support the point she’s making). Overall a very weak article, from a stats point of view. (And yes, I am amused at the fact that the judges seem oblivious to this.)

Lescroart - It seems all over the place, from a writing point of view; almost half the piece is devoted to a history to DIPS theory, which given the size of the article and the focus I think was way too much material and left almost no space to discuss FIP. It would have also been good for him to show how FIP-ERA holds up as a predictor of future talent level - again, a correlation coefficient or a MAE/RMSE would go a long way toward proving his point. (Irony point - for the second week, Lescroart uses a THT/Fangraphs stat. Hillaripus. Also funny is Will’s admission that he’s not familiar with FIP.)

I’ll do more later. I think so far the Knicker piece is my favorite.

Oh, hell, I misspelt Kniker there the second time. Deepest apologies.

Colin/#75

Don’t worry about it.  Typically more people misspell my name than spell it right.

Now for the three explicitly BABIP/DER pieces.

Hissey - Weakest of the three. He spends an awful lot of time discussing individual team DER as affected by specific fielders, but never provides any insight into how he came to those conclusions. With DER being such a simple stat to figure, he could have easily given a bit more meat to his analysis. To be awfully blunt, in two weeks so far Hissey has seemed content to simply use metrics to explain things, and he doesn’t come off as very convincing in knowing or caring about them any more than that. This probably comes off as unkinder than I mean to; not everyone that writes in this field is an analyst and not everyone needs to be. But if I didn’t know how to figure DER before reading this, I wouldn’t really know much more about it after reading the piece. And for the focus at hand, that’s a big weakness.

Euston - This is better and a big step up from his article last piece. It was clear and it explained the concept well, as well as showed a good understanding of the concepts involved. PADE was almost an afterthought, unfortunately - in 1,500 it’s hard to shoehorn a lot of things in there, and explaining PADE better would have probably meant explaining DER worse, which I don’t think would help any.

Swartz - The best of the three, and so far it’s my favorite article from Matt. It’s got enough detail to the numbers and the graphs without being overwhelming.

So far I’ll call it a tie between Kniker and Swartz for me. Both very good pieces.

One more aside: Goldstein’s dislike for individual fielding metrics again cracks me up, for reasons I cannot entirely explain.

I’ll comment as I read them…

Brian O’s article about first pitch strikes:
http://www.baseballprospectus.com/article.php?articleid=8931

I like the effort and the citing of Burley’s work, which is great.  But, it didn’t do anything for me.  Indeed, I imagine he’d get similar results if he looked at “second-pitch strikes”.  The problem is that the subject is rather complex, and it’s tough to follow on Burley’s work.

Tyler:

http://www.baseballprospectus.com/article.php?articleid=8933

Way, way, way too basic.  A decent article for a mainstream audience.

Bryan L (uh, how many Br*an’s are there in Idol):

http://www.baseballprospectus.com/article.php?articleid=8939

Very good handle on the whole thing.  If you read the comments by Will (non-BP?) and Christina (non-BP!!), you can see that the judges can’t decide what’s appropriate.  Since I don’t believe that BPro is the center of the universe, good for Bryan for using FIP, which is so simple, it should be a common metric for the mainstream.

He lays it out well, but he doesn’t sell it well enough.

Three more.

Knight - He never got around to explaining OBP or WHIP at all. We’re supposed to know what they are going in, I guess. If that’s the case I don’t know why we need an article about the basics. I’m bored here, very bored. Knight is at least good enough to acknowlege his mistake from last week without being defensive (am I the only one bothered by Will objecting to this?)

Oakchunas - Never really establishes that first pitch strikes are better than second or third or fourth pitch strikes - I think simply knowing that a pitcher threw ANY strikes is going to skew the results, and he never controls for that. Probably too much for a “Basics” piece that’s supposed to fit a 1,500 word budget.

Cartwright - Brian, as always, brings a gun to a knifefight. As always his analysis is outstanding, but he seems to be the writer making the word limit here work against him the most. It’s hard for me to say that’s an introductory piece about park factors. Again, it is very solid analytically.

I’m still going to have to go with Swartz and Kniker in a dead heat, with Brian very close.

I took Tango’s advice and did not comment at BP on any of my competitors articles.

I read them all, making notes - the fewer notes, the better.

Thumbs up
8 - Kniker, Lescroart, Euston, Swartz
7 - Knight, Funck
and myself, of course

No vote
6 - Oakchunas
5 - Hissey
0 - Ghiroli

Brittany:

http://www.baseballprospectus.com/article.php?articleid=8936

Sorry, she looks out of her element.  I like the effort, but she has too many conclusions, when really she should have alot more questions.

Tim - I (as you can see) have a pretty difficult name for people as well, so I know how it goes. Thanks for being understanding.

All in all I think this week’s crop of articles is better than last week’s, but again there were some real clunkers that felt out of place. So far I’ve come accross a handful of submitted pieces on other websites that assure me that this wasn’t due to a lack of better alternatives:

http://www.fantasypros911.com/bp-idol-entry-the-death-of-howe-news-bureau.html

http://frommomsbasement.com/2009/05/19/bp-idol/

http://www.beyondtheboxscore.com/2009/5/15/876896/playoff-probability-added

A question that has been bothering me for a while - if BP wants to get a more mainstream/reportery profile (the judges all seem to be in love with the idea of team reporting) - why not reach out to the dozens of baseball beatwriters and columnists that are being put out of a job as the newspaper industry retrenches?

Colin/81 - some comments on my own article

Goldstein didn’t like me picking on Olney. My first draft started with “How many times have you heard an announcer say there were 250 homers in this park”, and would base it on how I would write a letter to the announcer suggesting a better way to explain it to the radio audience. However, I wanted to give a concrete example, googled Yankee Stadium, and got Olney’s piece. It was just textbook misuse of numbers. After 555 words of talking about Yankee Stadium, I thought it would be a great piece for FanGraphs, but it would need to be a little closer to 1500. In my best attempts to stay ‘basic’, instead of groundbreaking analysis, I thought of how I would explain park factors to someone face to face (like I did when Jinaz came over to my house for the WS last fall), working in some of my own methods. I had three or four paragraphs, and juggled them around to get some kind of logical progression from one topic to the next.

Overall, I was happy with the result. Given the current BP readership, I don’t know if I’ll win, but I feel confident I’m top half of the original ten.

Matt S:

http://www.baseballprospectus.com/article.php?articleid=8932

Explaining BABIP.  Good, decent.  Not much more to say.

Matt K, OBP:

http://www.baseballprospectus.com/article.php?articleid=8938

It went way too long.  The OBP should have been a setup to something better.  I think he tried, but didn’t get there.

Brian, please don’t take my comments the wrong way - you’re one of the better analysts out there, and you deserve to win as much as anyone else in this contest (okay, so that’s a lie - you obviously deserve to win more than some of them).

But when you say that you wrote an article as though you were talking to somebody like jinaz, well that sounds about right. But jinaz is really freaking smart. And you can make a case for jinaz having a better background in stats and particularly baseball stats than 1/3rd to 1/2 of the BP Idol finalists, much less the BP Idol readers.

Really LOLing.

Well, jinaz actually was in my living room, and I was explaining how I do park factors, and a little differently from how others do them. Just spoken words, no charts, even if he is going for his PhD.

Brian, some of your park versions are a little off. Angel Stadium should start at 1998, Camden Yards should include 1992-2000, Coors Field should start at 2002 for the humidor or 2006 for the super-humidor, Dodger Stadium should start at 2005 when a huge amount of foul territory was eliminated, the Metrodome should start at 2004 (switched to field-turf) or start no earlier than 1994 (removed 6 ft. high plexiglass from OF walls), AT&T Park should include 2000 to 2003, Skydome should start at 2005 (switched to field-turf). Finally, I wouldn’t include the Washington Senators 1971 season at RFK.

I ran a quick correlation (min PA 150) on G/F and HR/F from ‘89 to ‘99. Needless to say they were (weakly) negatively correlated, about -0.28. I also can’t figure out how she’s figuring G/F; it doesn’t seem to match up with how it’s figured on Fangraphs at all.

terps - I used the KJOK parks db, with modifications as I became aware of new info or errors.

Angels Stadium - I have version 4 starting in 1997 with rf changing from 270 to 330. Did not create a new version for 1998, but apparently should, as rf wall went from 10 to 18.

Camden Yards moved home plate in 2001, then reverted in 2002, so v1 is 1992-2000 and 2002-2009, but that confuses the sql code when labeling.

I change Coors in 2005 for the humidor.

KJOK doesn’t have backstop distances for Dodger Stadium. v6 started in 2001 for cf 400 to 395. I will make a change for the foul territory.

Metrodome - I have the FieldTurf noted for 2004, but was still artificial before. I can run a test to see if it has meaningful changes. Didn’t know about the 1994 changes.

San Francisco - KJOK has the dimensions changing in 2004, lf 335 to 339, cf 404 to 399

SkyDome - again, changed types of artificial turf, I had it noted, can test

RFK - although 1971 was a different league, config was the same and results were within error margins of 2005-2007. Looking at it now, the changes from 1970 to 1971 were less than 2 ft, so may be insignifigant. I can test eliminating that change and running it 1963-1971 compared to 2005-2007.

Thanks for the feedback, it’s a challenge to keep up with the details.

Yes it is a challenge to keep up with all the park changes. I have found that it is sometimes better to ignore minor park changes.

But the switch from astroturf to grass-like turf is a pretty significant change. From 1999-2003 Minnesota’s run PF was 1.04, from 2004-2008 it was .95. Veterans Stadium was 1.02 from 1996-2000, and .90 from 2001-2003. Skydome was 1.07 from 2000-2004, and .98 from 2005-2008. Candlestick was 1.00 from 1975-1978, .94 from 1979-1981. I’d say the switch from astroturf to grass-like turf reduces run-scoring by at least 5%.

There have been no changes to AT&T Park since it opened. The distance marker was revised in 2004, but the fences weren’t moved. The same can be said about Dodger Stadium in 2000. The fences didn’t change. But some foul territory was reduced prior to 2000.

The LF alley was moved back 17 feet in Angels Stadium in 1998.

From a comment in Brian’s piece:

“This is Baseball Prospectus, not the FOX game of the week.”

Tim K, Run Matrix:

http://www.baseballprospectus.com/article.php?articleid=8929

Excellent.

Brian C’s piece blows everyone else away.

He not only tackled the basics of the issue, but he brought original research.

I thought this whole thing was kind of cheesy from the first I’d heard of it.

And now I see the headshots, and I realize there’s nothing “kind of” about it.

I’m going to give the BP voters the benefit of the doubt and assume that no one is basing their vote on a picture, but why include anything other than the writing at all? 

No offense intended to any of the participants.  Brian, you certainly look as if you have the most gravitas

Does anybody know what happened to Jeremy Greenhouse? He was one of the finalists, but didn’t have an article submitted for The Basics.

Jeremy submitted an audition, but after being selected a finalist announced that personal commitments would keep him from being able to continue.

In case Matt Swartz reads this:

If you are goiung to state that something is “clearly a dilemma” your following description of the situation should actually be a dilemma.  A Dilemma is when you have at two solutions to a problem, both of which lead to an undesirable result.

What you called a dilemma, otoh, was merely two statements that seemed true, but opposed each other.  That is conundrum, but not a dilemma.

Paul/100:

Thanks, I think I misuse that a lot.  I better be more careful-- I guess I’ll use conundrum instead in similar situations.

A nitpick with Brian’s article.  Where he says “that means there’s a 70% chance the ‘true’ value is between 1.30 and 1.60, and a 95% chance of it being between 1.15 and 1.75”.

.... you can’t make those statements with traditional statistics.  In fact, since the Yankee example was specifically selected because it was extreme, I think that from our perspective, there’s a much higher than 5% chance that the “true value” is less than 1.15.

"Angels Stadium - I have version 4 starting in 1997 with rf changing from 270 to 330.”

??? Typo?

Angel Stadium has never had a 270 foot right field.  I don’t think any team has had that short a fence in a very long time.

Looking at KJOK’s park database, Brian meant to say 370 ft. This is a mistake in KJOK’s database. It was never 370 down the RF line. It was 330 down the line and 370 to the power alleys from 1980-1997.

I obviously exaggerated when I said the switch from astroturf to grassy-turf decreases run-scoring by at least 5%. A more realistic guess would be a 2-3% decrease.

Has anyone else noticed how cynical the reader comments are for these articles (not on this blog, but over at BP). I have seen a lot of comments where the reader trashes the article and offers no constructive criticism.

DFL/102 - Can you explain your comment in more detail?

I started the article using Olney’s comments as an example of misuse of numbers in the mainstream press. I first thought of us saying “How many times have you hear an announcer say...” but I thought the specific example worked better. I pointed out that you can never extrapolate from a small sample - that’s a basic in any type of numerical analysis. When I wrote my article a month later, the numbers had moderated, but were still high. The chart was from a table I did last year, with 1, 2 and 3 year samples. I didn’t have time to construct quarter season samples. The table did show nicely how the errors got smaller with larger samples (another basic concept for any analysis). I offered the possibility that after a full season the ratio might very well be less than 1.45. Without going into a full blown explanation of standard deviation, I wanted to say what the numbers in the table meant in real life, and without using the term ‘confidence interval’, describe the practical application of those numbers, IF WE ASSUME that the 1.45 holds up until the end of the season. At that point, it was just for illustration. I held out the possibility that later on the new Yankee Stadium might still be average (close enough to 1.00).

There were several people, including non-BP staff that I respect, who thought the article was not basic enough for the assignment. We were instructed to model the articles after the series of ‘Basic’ BP ran five years ago, and a link was given to the archives. http://www.baseballprospectus.com/news/?column=31
I skimmed over most of those articles, and I honestly feel that my article was no more advanced than Silver’s “Science of Forecasting” (which featured four graphs on aging curves, seemingly the real subject) or Sheehan’s “Stolen Bases”. I also think that many who felt the article was not basic enough might not realize how much I didn’t put in, that I have already written thousands of words in several articles that go to a lot higher level.

Thanks to those here who offered help on ballpark configurations. I have been doing further research on the KJOK db on Wikipedia and other ballpark websites. I am focusing on 1998 and later, major and minor leagues, for my forecasting, and have recently completed the first table of all 300 parks listed in GameDay from 2006-2009, adding data from Wikipedia and Google Earth including latitude, longitude, home plate angle and elevation. I still have to do a yearly table listing dimensions, names and other things subject to periodic change. My intention is that GameDay’s play by play and the detailed parks db will result in vastly improved minor league park factors.

terpsfan,

The comments have a “pizza hut delivered the pasta” type feel to them.  If you take one or two of these articles and slap on Tango or MGLs name to them, readers would probably laud the efforts.  But instead, it’s free reign for the anonymous dickwad theory to take full effect where unknown commenters feel they can criticize writers “on their level.”

Nice food analogy Eric. I found this comment on Brian’s work ridiculous:

“‘Home run’ is two words; getting that wrong in your first sentence is not a great start. Getting it wrong again later in the article is worse; you can’t possibly sound like an expert at that point.”

Sometimes I write HR as one words, sometimes I write it as 2 words. Neither is wrong.

http://www.baseballprospectus.com/article.php?articleid=8934#commentMessage

The joker who made the comment quoted above received a rating of “-3”, which means that other readers didn’t care for his comment either.

Interestingly, none of the BP authors have a “rating” next to their posts (either as in the first 3 judges comments, or when they post later on).  You can see that while Sky gets a “-1” in reply to Will, Will gets no rating at all. 

Judge not, lest ye be judged.

I find myself not really enjoying any of Brian’s BP work so far. It’s the equivalent of asking Richard Hofstadter (when he was alive, RIP) to write a middle school social studies book when he’s already writing more entertaining and intellectually stimulating pieces.

Happy to be attacked by Will twice, once during Brian’s piece!! I started a nice little fight there lol.

Brian/106,

In general, you can’t say that a PARTICULAR 95% confidence interval is “95% likely to include the mean”, it’s that if you create confidence intervals for random samples, they’ll include the true mean 95% of the time.

When you’re looking back over the results for all these samples, if your eye is drawn to anything out of the ordinary, there’s a pretty good chance that that’s one of the samples that is in the 5% where the CI misses.

So if, at the beginning of the season, you’d decided to do these calculations on the Yankees new stadium, you could have said “there is a 95% chance that when I make my confidence interval in late May, it will include the true mean”.

But now ... given that the early data stands out as being so anomalous ... there’s evidence that we’re looking at one of the 5% exceptions.  This is based on the idea that we would have expected to see the new Yankee HR PF closer to 1.00 than to 1.45.  (If the new stadium was at 8k ft altitude or something, maybe it would be the other way around and 1.00 would look more surprising).

The basic idea is that classical statistics just doesn’t give us the ability to say “this is X% likely to be true”, it only gives us the ability to say “if ABC was true, we would see this result X% of the time” or “on a random sample, this process would work X% of the time”.

In this particular case, I don’t think this is a trivial semantics distinction either ... I’d guess there’s something like a 20%+ chance that the true Yankee HR factor is less than 1.15

First round results are in:

http://baseballprospectus.com/unfiltered/?p=1288

Lescroart gets bounced from the balloting - I understand the format means that there is no one single reason one person stays while another leaves, but it seems a very odd choice; Lescroart seemed very middle-of-the-pack to me, not one of the top-tier guys but not one of the actionably bad submissions of the batch. Hopefully he decides to continue plying his craft elsewhere - either at his own blog or a place like BtB - and can hone himself.

And congratulations to Brian Cartwright for having enough people not listen to Tango.

Agreed with Colin.

So are they keeping Byron around? Sure didn’t sound like it from the article or the comments, which makes Tango’s idea that they’d keep all ten go right out the window. I don’t think they’ll keep any of them, other than the winner and even then, I doubt they’ll keep them beyond the timeframe they listed. I see that Sky is already recruiting, which seems a bit bushleague.

Kara said:

which makes Tango’s idea that they’d keep all ten go right out the window

I said:

Put them out of their misery Will… when it gets down to the final 5, hire them all, and announce it.  Split the 1000$ among all 5.
...
The reality is that at the very least, three if not five, of those people will be producing regular or semi-regular freelance work for BPro.  If we accept this as the eventual reality, exactly why are we making Brian and the others do what they are about to do?

I did say that they COULD accept freelance submissions from all 10 of them right now, not that they WOULD.

This, from Will Carroll’s twitter feed, might give some idea as to where they’re headed:

“I love baseball, not numbers and equations. I’m stunned at the people that prefer the latter and that think it’s a biz model that works.”

The biz model that works the best is the Fantasy model, which is 100% numbers-driven.

BPro enjoys being thought of as sabermetrically-driven.

Puck Prospectus, recently launched, now in the hands of Will Carroll, is numbers-driven.  Are we supposed to see a bait-and-switch?

"I see that Sky is already recruiting, which seems a bit bushleague.”

What do you expect from “lessor sites”?

The letter from Will informing people they became a finalist stated that after seeing the auditions, BP may in the end hire 3 to 5 of the 10.

Kara, who says I’m recruiting?  I want to get in touch with the guy.  He’s never published anything else on the web and I can’t find an email address for him.

Unlike BPro, I can spot talent on my own, without self-indulgent gimmicks.

I think a lot of the disgust with and vitriol toward BP recently comes from their move from being a sabermetric site to being more like ESPN Insider or other mainstream baseball reporting and commentary sites.  Those of us who loved reading stuff from Keith Woolner, Clay Davenport, Nate Silver, and Dan Fox are disappointed to see BP go from a place where the people who were mocked for loving baseball AND numbers could gather and read about the game from people with a similar bent to a place where people are mocked for not getting their noses out of their spreadsheets. 

You’d think Will would see the irony in that even if BP’s dramatic shift in tone is a good business/marketing decision (which I’m not sure about, given that the typical fantasy player, which they appear to be targeting, loves baseball AND numbers even if he doesn’t understand the numbers very well).

Unless BPro believes that stat analysis articles will drive people away, I don’t really understand the “business model” rationalization.  If they were to let Brian write about hard-core sabermetrics for them, that doesn’t mean that Carroll/Kahrl/Sheehan/Goldstein are going to have to write less articles.  It just gives them a more diverse voice, while not alienating their original core audience.

I’m sure they could find someone of decent quality who was willing to write for free just to get their work exposure.

This whole Idol experience has just served to clarify and highlight the shift in culture over there.

Some of us don’t have a business model, and just think that if you have baseball numbers and equations that they should be correct.

No wonder they didn’t hire you.  You care more about the truth than making money.  You wouldn’t fit in there at all.

Does anyone else think that those in charge of BPro might literally believe Huckabay’s “sabermetrics is dead” screed, and thus be angling BPro away from being at the forefront of sabermetric research?

Congrats to those who are still alive in the contest.

Matt #126, I’m going to take your last sentence as a reference to your first paragraph, which makes it much more amusing for me.

I want to reiterate my point that I made in the other thread about this-- I think BP is using this contest to see what readers want. 

I’m surprised that more of the comments in these threads aren’t finding their way over to the comment threads on BP.  As a contestant, I didn’t want to say this before because I think it would sound underhanded (and probably would have been), but I was really confused why no one pointed out that Lescroart used 2007 pitch f/x data for Hafner as people mentioned a lot over here.  That’s a really big mistake and no one mentioned on the actual article.  The readership of this blog seems to want BPro to take a more saber-oriented approach, and many of you still have subscriptions.  Why not call people out?  Brian’s article this week was fantastic, and I would have expected to see some of the positive comments from here pushing him forward over there.  I really think BPro is testing what people want to read, and it seems like people here have strong opinions on that.

Matt S., I’d be curious to see a roll call of people here on whether they have BP subscriptions or not, if they’re willing to say.

Mine expired in April 2009, and I’d had it since they started subscriptions in 2003, if I remember correctly.

I’ll put up a poll, if you give me a minute…

Poll is up:

http://www.insidethebook.com/ee/index.php/site/comments/poll_do_i_subscribe_to_bpro/

Mike/#127, That was a bit unintentional.

"I think BP is using this contest to see what readers want.”

What if they want wwtdd.com? (NSFW)

--- --- ---

One of the judges mentioned that some articles receives many more views than others.  I’m surprised we haven’t discussed the ballot-stuffing potential for this contest.  Or, perversely, the potential for a Vote For the Worst campaign.

The ballot-stuffing will be impossible.  It’s 1 vote per subscriber.

Tango 134/ There’s got to be some way to get around that.  Maybe using a googlebot.

I don’t think ballot stuffing is a realistic option in this case, because of the subscriber issue.

But it’s really 10 votes per subscriber, as everyone can vote for as many entries as they want. (Yes, only one vote per entry - but the more entries you vote up, the less any of your votes are worth.) Also, this comment from Will Carroll was very enlightening:

“That said, it’s the raw number of “thumbs up” votes that decides who stays and who goes.”

I don’t see why that should be the case - it should either be, in my estimation, number of yes votes minus number of no votes or a percentage of yes votes of total votes (and we know at least one of these methods would have kept Lescroart in the contest.)

Right, the raw “yes” vote way is wrong.  You have “yes”, “no”, “abstain”.  If you have the following:

Writer1: 100,20,600
Writer2: 110,60,550

Why would Writer1 win?  That would be identical to counting an “abstain” as a “no”:

Writer1: 100,20+600
Writer2: 110,60+550

This is a counting v rate stats debate.  The counting side usually loses.  I would have gone with a yes minus no, or give 2 points for yes, and -1 for no.

Right, ballot-stuffing is not possible.  Since they know exactly who the voters are, and how many votes per entry they are selecting on, they can’t ballot stuff.  It’s totally impossible to break that one.

This is the way it works programmatically:
If entry_switch = ‘T’ then 1 else 0

So, there’s nothing to hack on the database side.  The code doesn’t need to run on the server.  So, 100% foolproof.

But in the commercial world abstain and no ARE the same.  All you care about is who buys your product.  If a “Yes” signals a desire to read more articles like this one, and BP want to please (and add) subscribers, then it’s not clear their method is the wrong one.  The only reason to give extra weight (negative) to No votes is if you think such articles will literally cause people not to subscribe. Otherwise, indifference and dislike are functionally the same.

I bet if you look at the voting totals versus the comment totals, you’d find a correlation. (If Eric or any of the other BPers are auditing this thread still, would you look into that?) My theory is that as articles became more commented on they became the more popular selections to read, regardless of what the comments said.

The other two ideas I have for massive vote total discrepancies between articles are outside referrers or a problem with the randomizer. Both problems are magnified by the raw yes vote method.

One has to wonder what the purpose of the “No” option really is, then.

Guy, right, what Ryan said. 

Again, is 100 yes, 0 no, 500 abstain really worse than 200 yes, 200 no, 200 abstain?

The latter leaves you with a bad taste in your mouth that would cause you to not subscribe.  The former is a niche writer, who delivers the goods to those who bothered to read him.

I exaggerate the illustration, but seeing how close the votes are, I wouldn’t need much exagerration.

Voting is open to everyone, right?  My point was that if Tim Raines (see what I did there?) were a contestant in BP Idol, he’s go on Ellen and Larry King and tell people to go to BPro and vote for him.  Writers currently with large audiences will encourage readers who may not not have a current BPro account or an interest in voting to go vote.  For them.

Sky, voting is not open to everyone, only to people with subscriptions.

I have gone on Facebook and and linked to each week’s main page, asking my friends to read and vote, knowing only subscribers can vote, but many of my friends are subscribers, if not staff. Family and folks from church are much less likely to be subscribers, but they could reead the article of they were curious. I could spend $4.95 a person to get each of them a gift subscription for a month, but without knowing the total number of votes needed to make a difference that could add up quickly.

Yes, voting is only open to subscribers.

Which leads into Guy’s point - BP has no way of knowing which of these readers will lead to ADDED subscribers, based upon votes, because they’re deliberately sampling from subscribers only, and it’s very possible that non-subscribers have a different set of preferences (I’d even consider it likely).

So essentially what you’re measuring is the preferences of:

1) People who will subscribe regardless of outcomes, and
2) Leaners, people who are considering renewal but aren’t sure either way.

For the second category, the identity of the winner may sway their choice - if you like reading one writer but not another, you’re more likely to renew if the former is voted in than the later.

I think net votes or a percentage is the correct choice there.

We know this week the margin was under 10 votes. So max of $44.55 for Lescroat to have bought himself entry into Round 2.

The early rounds seem to favor people who have large friend networks/name recognition among BP subscribers; it seems as thought the majority of voters were unable to read all articles. That’s about as large a “ballot-stuffing” factor as you’re likely to find.

I sat down at the computer Sunday night and asked my wife not to bother me (which would interrupt my concentration). Took me about 2.5 hours to read the other 9, 15-20 minutes each to read, ponder and take notes. At the end I gave them numbers 0-9 and decided who to vote for.

The problem with a percentage approach is that it encourages “bullet voting” by friends/supporters of authors.  If I like Brian, I vote yes for him and “no” on all other competitors every week.  If even one author does this, everyone else kind of has to do the same to have a chance. 

If you want to measure intensity of support—and I agree that’s a valid goal—then maybe you let voters rank their top 3 or 5 choices, with more points for a higher rank (like an MVP or CY ballot). 

As it’s set up now, I don’t know how much information is in the No votes.  Maybe net voters is a better measure of quality than total Yes.  I think it’s hard to know....

Here’s how I would handle the balloting:

* Users rank every contestant from best to worst. They can rank as many or as few candidates as they like.
* Assign each candidate a number of points based upon where they rank.
* The total number of points per ballot is determined by how many entries are on the ballot.

For instance, if I simply rank one entrant, that entrant gets one point.

If I rank four entrants, my top-ranked entrant gets four points, the next three, the next two, etc.

If I rank all then, the top-ranked gets ten, the next nine, etc.

That way you allow people to rank as many or as few as they are comfortable with, without rewarding someone for voting for themselves to the exclusion of others.

Colin I think that is very well thought out and would work well, but I think BP wanted to make everything identical to the TV show. In a comment where I made a suggestion, Will asked me if I watch the show, I said “No” and then he suggested I did so that I would understand BP’s contest.

This discussion reminds me of Arrow’s impossibility theorem (that is, no voting system can always fairly reflect the will of the public as a whole): http://en.wikipedia.org/wiki/Arrow’s_impossibility_theorem

I very much like Colin’s suggestion.

It’s been bugging me for a while, and I finally tracked it down. What I just suggested is apparently called a “modified Borda count.” I have no idea where I originally encountered the idea.

Colin/#150

I too like the idea that you presented, though I can imagine the howling.  I can see someone who doesn’t fully understand the process to complain that if they only rank 1 person, they FEEL that their vote is only worth 1/10 as much as someone elses.

I think in the grand scheme of things, the best would have been a Yes/No/Abstain type approach.

My one hope is that a lot of this complaining will go away once the readers/votes realize what the rules are in relation to them.  Basically,

—We are counting YES votes only
—You will only be given a couple of days to read and absorb all the articles.

If there is one thing that can be complained about is that the rules haven’t always been consistent or clear.  In just our last instructions we were told about proper table formatting and they wanted to have the articles in standard format, but they really haven’t said what standard format is.

Week 2 articles on Fatnasy are up.

http://www.baseballprospectus.com/bpidol/

Brian,

If you read this—How big was the sample size you used? Dates?

Thanks

After reading about half of this week’s articles, I don’t think Brittany Ghiroli really knows how BABIP works. Nor do I think she’s ever looked at BABIP on a player card other than Crawford and Melky.

Keep in mind the league average for Babip hovers around .300, so if a guy is posting a higher average than Babip, it’s likely his average will go up more as his Babip increases.

----

Crawford’s Babip is above his average, which means he’s finding a lot of holes and figures to come down some.

Ummm, what?

I did really enjoy Matt’s and Brian’s articles. Brian’s article was a big step forward from last week in terms of how it was written, despite the lack of research (and despite the fact that he’s never played fantasy baseball). That’s not at all a criticism--you can’t really do a whole lot of research for this kind of piece.

Ken Funck’s article had almost nothing to do with fantasy, but I enjoyed reading it (and I’ve never played Strat).

First thing I need to do this week is start writing before 6pm Thursday. The plan was to think about it on Tuesday, write the first draft after work Wednesday night, then consider and revise it Thursday night. Come Wed, I was tired as crap after work and couldn’t think straight, so writing didn’t get started until lunch break on Thursday. The articles didn’t finish as well as I would want without the time to reflect and then go back to wrap it up.

I’ll comment on the order I read ‘em. 

Matthew K:
http://www.baseballprospectus.com/article.php?articleid=8973

Alot of “I think this”, and “I think that”.  No basis really.  No advancement in knowledge.

Tyler H:
http://www.baseballprospectus.com/article.php?articleid=8977

Decent data, actionable items.  As other readers noted, I prefer substance to style.  And whatever misgivings you have on style here, the substance makes up for it.  Decent article.

Cot’s Contracts:

http://www.baseballprospectus.com/article.php?articleid=8976

Good information… for a GM.  It’s the kind of article that has good resources for analysts, but I see nothing in here for Fantasyers.  He brought a knife to a gun fight.  But since I’m wearing a bullet-proof vest, Jeff’s article appeals to me, even though it won’t appeal to most others.  Good reference material, which is what I want.

MLB.com reporter:

http://www.baseballprospectus.com/article.php?articleid=8975

As noted by others, not a good enough understanding of BABIP.  Again, not my cup of tea.

Brian C:

http://www.baseballprospectus.com/article.php?articleid=8972

Easily the best so far.  It’s not even funny.  Any criticisms is really being nitpicky, within the constraints of the word count and contrived topics.

Ken F (Strat):

http://www.baseballprospectus.com/article.php?articleid=8970

Enjoyable read, but, really, nothing else there.  Ken was a victim of the contrived theme topics.

One thing I noticed is that the names of the authors are impossible to remember, week-to-week.  The names are simply too similar, which is why I highly prefer internet handles.

***

Tim K:

http://www.baseballprospectus.com/article.php?articleid=8968

Decent article.  The “percentage” method is ok, but not as good as the marginal change divided by standard deviation.  I suppose it’s a good way to introduce someone into the better version, similar as RC would lead to BaseRuns.

And Will’s complaints are starting to border on the caveman’s “your technology confuses and frightens me"… just as he is about to talk about MRIs and CAT scans!  If Will Carroll is a target audience of BPro, then there are at least two distinct audiences being catered to, and there should be two sets of judges and two competitions. 

I may as well argue that Will’s articles lack any sabermetric principles whatsoever.  See?  It makes no sense at all, since Will has a different target audience.  Will is projecting HIMSELF here, when he really should evaluate based on the target audience of the article, NOT the target audience HE wishes for the article.

Matt S (statspeak.net):

http://www.baseballprospectus.com/article.php?articleid=8974

Very good piece.  Matt and Brian are the leaders and should simply be hired right now (along with Colin).  Really, anybody at statspeak.net.  If they’re good enough for Pizza Cutter, they’re good enough for anybody else.

Boy, that Will is so incredibly defensive of anything BPro.  In both Matt and Brian’s pieces, Will puts a shield around PECOTA.  Given that Will admits to not follow the heavy-mathy around anything, he’s the last guy to defend PECOTA on the merits.  He can’t have it both ways.

BPro Derek Jacques says:

...it’s pretty unfair to accuse BP of not being “aggressive enough” in improving PECOTA, given that there have been improvements made to the system just about every off-season since it was first introduced.

Wrong.  It is very fair to criticize any black box system that produces results that need further explanation (which is not forthcoming).

Brian O (VORP, VOFP):

http://www.baseballprospectus.com/article.php?articleid=8971

He’s in the same league as Tim K in terms of the evaluation.  He uses (seemingly) some sort of system I use, but ignores positional scarcity?  Plus, he tailors it to some league that really has little to no literature on it, and therefore, requires too much effort on my part to validate his various claims.

Too early to tell.

I like Timo Seppa’s criticism:

My biggest qualm: “It goes through a hundred years worth of baseball players” - That is not true and is a basic fact that should have been checked. From the BP Glossary: “PECOTA compares each player against a database of roughly 20,000 major league batter seasons since World War II.” That would be 64 years. As good as the article is, you can’t make mistakes like this. Thumbs down.

This from the guy who’s talking about ten and thirteen game sample sizes from goalies and calling them reasonable.

Really, you can’t make mistakes like that?  Thumbs down?  Wow, pretty harsh.

And yes, Timo made a seriously egregious error in his article regarding sample size.  Matt’s error has zero impact on the actual article, and only has impact if you are trying to infer his research thoroughness.  That is one huge nitpick.

“Very good piece.  Matt and Brian are the leaders and should simply be hired right now (along with Colin).  Really, anybody at statspeak.net.  If they’re good enough for Pizza Cutter, they’re good enough for anybody else.”

Tom, I think that’s the nicest thing you’ve ever said about me. 

Yeah, I thought it was best to not respond to that comment.  It’s funny that guy was already known for making sample size errors and stuff like that.

Matt, I hadn’t read them all yet, but yours was the best I’ve seen so far (even better than mine).

I was comparing your numbers to Oliver’s predicted BABIP, for all I fell between you and PECOTA, using only historical data, no comparables.

BP Me You
334 321 306 Soto
307 295 286 Francoeur
300 308 324 Castillo
323 332 342 Young

Thanks.  I liked yours again, of course. 

I knew that batted ball article would end up being relevant to some topic, and so I hope I didn’t peak early.  I’m still waiting to get ready on the next topic.

If you want to check your BABIPs against the rest of mine, they should be copy/paste-able from the StatSpeak article, and if not, I can email you them.  It’s still kind of early to tell how the model does.  CHONE is doing really well, actually, I think better than my correlation and RMSE and better than PECOTAs too.

I thought your article was really good this weak Brian.  I don’t play fantasy, but as a Cardinals fan I am very interested in whether Ryan Ludwick’s season was “for real” last year.  So thanks for ruining my day

Anyway, I was shocked out how ridiculous some of the criticisms were on your article.  You definitely “improved” by dumbing down the stats, but I still read some really ignorant comments.

One person criticized you for using wOBA, then suggested that you should be a research assistant at BP “pulling numbers and getting things ready for other writers.” What an ass.

Expect Ludwick to fall back in his BA, due to an abnormally high BABIP for him. His projected BABIP has crept steadily upward over the years, mostly in 285-298 range, was 301 for 2008 314 for 2009. I think 301 is more realistic.  I don’t see any reason that HR%, BB% or SO% will change. I think he has a better chance hitting .270 than .300, but still with 30+ HRs.

Oliver Projections

2009 lo-hi 2001-2008
314 262-298 BABIP
077 058-068 HR%
081 072-083 BB%
244 229-265 SO%

jtrichey:

I think this author is a great statistical mind. I think he would be a much better research assistant at BP--pulling numbers and getting things ready for other people’s writing--than he would be a writer for the site himself.

Yes, yet another a$$hole comment. 

You have this intelligent comment:

On the negative side, there are a handful of typos and a few awkwardly phrased sentences, but that shouldn’t matter too much, as under “normal” circumstances, he’d have editor to clean that up.

But these comments are laughable:

That being said, I won’t be reading past the introduction on this one. It is a flood of errors and awkward phrases that add up quickly and make it hell to get through. Simply put, if the author can’t be bothered to ensure a minimum level of readability in his very first paragraph, then I have little interest in giving him a chance. Sorry.
...
If someone doesn’t pay attention to their writing style, I question how well they can analyze stats.
...
The first paragraph is terribly written. I had to shake my head a few times to make the words/numbers fall back into their correct places. The rest of the piece doesn’t get much better with a clear lack of succinctness - a professional writing course might serve Brian well. Thumbs down.
...
Perhaps it’s the English teacher in me, but I can’t get past that first paragraph without the urge to just draw a big red X over it, hand it back, and say “try again”! Honestly, does the author know anyone who has a grasp of basic grammar? I realize this is not a grammar contest, but it IS a writing contest.
...
I had the same issue, though, with the very first paragraph that you did, to the extent that I stopped after the intro, and only went back to complete the article later.
Upon doing so, I looked more favorably upon his work, but still could not bring myself to offer the thumbs-up. Perhaps next week.
...
The first sentence is tortuous, the second is meaningless. I care about grammar, and any time I read published work with this many errors in the first couple paragraphs, I stop reading. I scanned the rest of the article, and didn’t find enough to outweight my initial impression.

I think BPro needs to take off the ridiculous word count shackles.  Brian’s writing has been more than fine at statspeak and Fangraphs.

Really, these particular BPro readers are so being so overpicky (not an english word?  sorry about that… fail me.. should have said English too) it’s ridiculous.  Brian being forced to go through this hazing ritual is an embarrassment to BPro.

Tango/178- I completely agree.  Eric mentioned the “anonymous dickwad theory” earlier in the post.  That is clearly what’s going on here with Brian and I know that it is something that has plagued Dave Cameron at FanGraphs.  You have to imagine that some of the guys commenting were ones who’s articles didn’t make the cut, so they are just pissed.

Bergstrom talks too much.

Yes, there were several paragraphs pulled for word count.

I think some of the comments made here and over at BP this week reflect an ignorance of the conditions under which the contestants are writing and show two very different viewpoints on how to judge BPro Idol. (Pardon me, but deliberate American Idol references ahead)

On one hand you have the Abduls, who look at the writing style only and don’t bother to examine the content, either because they don’t understand it or because they don’t care. You could also call them the Carrollites.

On the other hand, you have the Jacksons, who judge on the substance and can overlook the writing flaws in favor of interesting perspectives and new ideas for the field of sabermetrics. You could also call them the Goldstienians.

In both cases, each expects a lot to be achieved within a four day writing window. For someone who’s going to be posting every day on BPro with instant analysis like Joe Sheehan or news reports like William “Paula Abdul” Carroll, judgment based on writing style is fairly reasonable. But I am certain that some people in the competition, if hired by BPro, would write on more of a weekly basis because of the research typically involved in one of their pieces. That makes the conditions they are writing under less than ideal, and can lead to things like using a different, possibly less than ideal method in an analysis.

Most of us here, including myself, want to see the analytical minds win out. But we can’t judge the reporters like we do the analysts.

Brian commented “wOBA is widely accepted at other sites that publish this kind of material.”

I’m trying to limit my snark in general, especially over at BPro, but I will say this… You mean there are other sites that publish this kind of material?

Matt, the reporters were asked for the first two weeks to provide analysis and for the most part provided analysis. They are being judged on this in a competition with analysis. That’s how they should be judged at the time.

Now, I would agree if you suggested that it’s an odd competition that asks reporters to analyze and analysists to report. I think they would have been better in a format that allowed each writer to do what they were good at.

And I think it’s really insulting for BP to go around looking for reporters this way, when there are obviously astoundingly qualified reporters/beatwriters/etc. available due to the newspaper industry basically failing. Carroll and Kahrl are both BBWAA members, right? Call the chapter president, get the names of a few people who just got laid off and do some job interviews if BP really wants more of a “reporting” profile.

Haven’t had a chance to read everything yet. I sent Kahrl an e-mail in response to a comment she made to Kniker’s article - basically just links to my SQL tutorials - but no response in two days.

My scorecard.

First column is the total.  Next is the name.  The final two columns is how much of a cup of tea each of their two articles was for me.  Ratings are 0-4 (0=not my cup of tea, 1=decent, 2=good, 3=great, 4=I wish I did it)

6 Brian C 3 3
4 Tim K 3 1
3 Matt S 1 2
3 Jeff E 1 2
2 Brian O 1 1
2 Byron L 2 x
2 Ken F 1 1
1 Tyler H 0 1
0 Brittany 0 0
0 Matt K 0 0

I think they should stop the competition when there is 4 or 5 people left.  The top 4 should be hired right away, as each would bring good quality to the table.  I mean, NO ONE does what Cot’s Contracts does, and he’s not going to break my top 3.  But, he’s valuable.

I really hope BPro realizes that they have ALREADY accomplished their objectives.  To force these fine people to continue with the never-ending hazing ritual, of subjecting them to contrived topics as well as some of the most nitpicky comments that would only make an editor proud, is insane.  Just admit that you will hire 3-5 freelancers, stop the competition when there is 3-5 people left, split the prize money, and pat yourselves on the back for a job well-done.

Otherwise, to be a slave to the process “just because” is being nothing more than a fool.

Seems like the hazing ritual is working just fine.  It’s bringing a lot of attention to the site, especially around here.  I’ve probably read more BP thanks to this contest than I have in the last 5 years.

If the top 5 get together they could always set up their own website and write the pants off the current BP powers that be.

Hazing rituals, like car wrecks, always get attention.

I guess that’s my question - why are you guys spending so much time talking about this again?

The lessons learned from this - the guys at StatSpeak are good enough to write for BP right now, BP has some really pretentious, annoying subscribers, and Will Carroll is remarkably ignorant about anything that doesn’t promote Will Carroll. 

We already knew all that.  Having Brian or Matt end up behind the subscriber wall isn’t even a good thing. 

This whole thing is dumb, and like most of BP, it should just go away.

The thing that BPro has is “branding” (or exposure), plus they pay their writers.  Apparently, that seems to be valuable to Brian, Matt and others.

Fangraphs for example doesn’t have the brand, but their exposure if pretty high.  They also pay their writers (not as much, but a decent deal nonetheless, especially without all the b.s.).

I guess in the end, we prefer highlighting the good works of others, and if that means taking down “the man” a notch, all the better.

Agreed with Dave.

The good to come out of this: I’m visiting StatSpeak more often, which had been off my radar for whatever reason. 

Probably not what BPro wanted, but it’s all I get out of it.

Yes… my evil plan is working!

I really want to hear if Brian feels like I do about this (and for that matter, Eric, since he writes for BP), but I think the criticism is very overblown.  For one thing, I don’t really feel hazed.  I feel like I’m getting a flood of readership, and more so than if I just wrote regularly for BP.  I don’t feel hazed-- rather, I find this challenging, and I like being challenged.

I’m not really sure what you meant by this comment, Tango/189: “The thing that BPro has is “branding” (or exposure), plus they pay their writers.  Apparently, that seems to be valuable to Brian, Matt and others.” Is the implication that we shouldn’t value being paid?  I’m not quite sure I understand you correctly and it is probably just bad wording, but if it were bad to prefer being paid to not being paid, I’d rather like a free copy of your book. Could you clarify?

As for the readership aspect, I’m really enjoying getting a widespread group of BP subscribers reading my stuff.  While I’m incredibly happy to be invited to write at StatSpeak, it doesn’t have the readership that BP does.  I spend a lot of time doing what I think is good research, and I don’t see it linked all that much even when people discuss topics I’ve looked at extensively. 

Take a really good example-- Dutton and Bendix wrote the article “Batters and BABIP” on THT back in December.  I’d been doing research on batted ball types and BABIP at TheGoodPhight in small doses for some time.  They wrote a good article and they got recognition for it.  It was a “post-dicting” article, describing BABIP luck in the past and positing what it “should have been.” I feel that it’s more useful to do predictive models of BABIP, and continued to do some research on that, publishing an article at TheGoodPhight on the topic a short while later.  I introduced a few new variables to predict BABIP in the future.  TheGoodPhight has a decent readership, but it’s mostly Phillies fans and not hardcore sabermetricians, so when Carty wrote his “What’s the best BABIP predictor?” article on THT a few weeks later, Dutton had apparently updated his model to include some of the variables that I had introduced in my previous article.  It was also now a predictive model, like mine was, though my model was not linked.  That article got forwarded around the net.  Since then, I’ve written two more articles extending the study at StatSpeak and there has been some discussion of them elsewhere.  But there are now 88 people who have commented on this article on BP discussing this research.  Yes, some of it is silly comments, but it’s readership of something that I have been pushing at the frontier of for a while now with little readership.  Yes, there is a subscription barrier to reading articles at BP, but there’s also an awareness barrier where lots of intelligent baseball fans just getting into sabermetrics haven’t bookmarked a million websites.  I don’t really understand the negativity here in implying that I am wrong to want readership or money for my work.  I’m not quite sure who the victim is here.

"You mean there are other sites that publish this kind of material?” (I may have missed out on the level of sarcasm in the comment)

THT, Baseball Analysts, FanGraphs, StatSpeak, etc publish statistical research articles like the ones I like to write, and I believe most of their readers understand wOBA. FanGraphs lists it on their stat reports.

“You mean there are other sites that publish this kind of material?” (I may have missed out on the level of sarcasm in the comment)

THT, Baseball Analysts, FanGraphs, StatSpeak, etc publish statistical research articles like the ones I like to write, and I believe most of their readers understand wOBA. FanGraphs lists it on their stat reports.

Hi Brian, I’m Sky.  I’m involved with Beyond the Box Score (see link in name), a site that tries to write articles like the ones you succeed in writing.

(Sarcasm aside, continued good luck to you in the contest.)

Yes Sky, I know who you are.

My comment you quoted was referring to stat analysis, I just wasn’t sure if you were referring to the same thing, or to the entirety of what was being published in the Idol competition.

I guess I wanted to be just like Dan Fox. He was my favorite writer at BP before he left last year to take a position with the Pirates. Then Dan gave my name to another team that was looking to hire a stats guy. I applied, they didn’t hire me, but as far as I know they didn’t hire anybody yet.

Matt: I’m not saying it’s wrong at all.  Some people value the “branding” of BPro more than others.  As for getting paid, money is money.  No one is going to begrudge anyone there.

So, I have no issue with you or Brian in terms of your career aspiration.

As long as you don’t think it’s hazing, that’s fine.  It looks like hazing to me, but that’s how I see it.  BPro readers should basically be super thankful they get to read your stuff.  Almost all the criticism are so picky, and they only exist because of the “Idol” thing.

I think we all agree that the competition is silly, in that BP can’t figure out what type of writer or articles it needs or wants, is basing its decisions on this contest instead of the larger body of work of the authors, and that 2 of the 3 judges aren’t statheads (though those same 2 may be the best assets BP has going for it these days).  And there are some who probably think this contest is beneath them and that BP doesn’t deserve to have them or MS or BC.  But if the end result is that BP gets a good writer that they wouldn’t have otherwise, and MS and BC get more exposure or readers or money--that’s not all bad.

I was not originally going to enter, but at least three members of the BP staff (not any of the judges) and others encouraged me to so I did. I do admire BP’s ‘brand’ and their compensation per article is more than I have been making. Even if I lose, I can still say “I wrote for BP.”

This is the first time they tried something of this sort, and like anything it might not have come out just the way BP expected. I understand Tango’s concerns, but overall I have enjoyed the process and don’t get bent out of shape at stupid comments. I have been taking this as a learning experience, and along the way have discovered that many of those that I admire in the community think well of me, and that means a lot to me.

Cot’s is out:

http://www.baseballprospectus.com/unfiltered/?p=1295

As I said:

Good information… for a GM.  It’s the kind of article that has good resources for analysts, but I see nothing in here for Fantasyers.  He brought a knife to a gun fight.  But since I’m wearing a bullet-proof vest, Jeff’s article appeals to me, even though it won’t appeal to most others.  Good reference material, which is what I want.

So, it’s a silly setup that would make it so that Jeff would be eliminated.  By this process, Doug Pappas and Will Carroll would never have been hired by BPro Idol.

I see Sanjaya is still in there.  You can probably guess who I mean.

Week 3 theme: “Baseball Below the Big Leagues”

Tango/#199, I have to disagree with you on Doug Pappas. He still had some good analytical chops, and I think could have been hired by the BPro Idol route. I pretty sure Nate referenced his intial work when building MORP. I’d check, but for some reason, BPro isn’t loading up for me at the moment.

I’m interested in what Matt and Brian will write about this week. Seems the only statistical topic regarding the minors is MLEs, and Brian’s already written an article about them so far.

Sometime tonight I will be putting word to pixels...I plan on writing about my experiences at the AAABA National Tournament, one of the premier amateur tourneys in the country, and how working there and in summer leagues in general gave me insight into statistics. Probably no numbers, just story telling, but hopefully some ‘take-aways’.

By the way, I want to congratulate Matt for having the most popular entry last week.  I think it is both a sign of the quality of his work and an indication that the general BP readership may be interested in more serious analysis.

I think blackadder’s missing the point. BPro has about 20000 subscribers right? Id guess that each of these is getting around 1000 views if that. voting is probably even more segmented than comments. bpros comments are a joke because the unemployed clowns like richard bergom are there a million times but its the same people over and over. most of the general audience that is coming over from espn isnt voting or even reading this.

im also curious if euston is getting hired by bp. of all the writers, he’s the one with something valuable (cots) to bring over. if they dont hire him or try, they won’t with anyone. if i were him, i wouldnt go. cots would be better off at bref.

I’ve got a couple random questions that, depending on their perceived interest in the community and whether other analysts have already addressed my questions, could turn into columns in the contest, so this seemed the best place to put them… Coincidentally, they both involve consistency.

=============================

The first: How much greater of a win value is associated with an otherwise average player who has a +10 defensive run value versus an otherwise average player at the same position who has a +10 hitting run value?

A basic RS vs RA Pythagorean-type simulation shows that a run saved really is more valuable than a run scored, but I think the effect is more than that. I’d hypothesize that defense is more consistent than hitting (i.e. the +10 defensive contribution is more likely to be spread out across more games than the +10 hitting contribution is). This is at least in part because balls in play are a more random event than hitting, since I’m pretty sure hitters usually accumulate run value against bad pitchers, which removes at least in part the randomness of that accumulation.

If that is indeed true, I’m guessing randomly knocking 1 run off of the opponents’ portion of final scores gathered from an entire season (on 10 separate occasions) will lead to more win value accumulation than non-randomly adding 2 runs to a teams’ own portion of the final scores on 5 separate occasions. (I know this not exactly how this stuff works, but hopefully it gets the basic idea across.)

Furthermore, it seems to me that there would be some sort of synergestic mechanism of gathering a bunch of defensive aces to amplify this effect, which might be an additional consideration when figuring out why teams who improve their defense considerably exceed expectations (if that is indeed true—I’m not sure it is—but it seems like this sort of thing might be at least part of the Rays success last year).

===================

The second: Is it possible identify streaky hitters and consistent hitters who will continue to be streaky hitters and consistent hitters, respectively, in future years?

This question is prompted by Jayson Werth proclaiming himself a streaky hitter. Assuming he has been “streaky” in the past, is it really more likely that he’ll continue to be “streaky” in the future than a randomly selected player? How about versus a “consistent” hitter? Defining “streaky” and “consistent” is tough, since it probably relies on manipulating game logs by figuring out inflection points or somesuch, I presume…

===============

If there’s material out there that’s already published regarding one or both of these questions, I’d love to see it.

Week 3 articles are up. If you need the link, see comment #156.

Immediate thumbs up from me for Kniker and Swartz. There are two or three others that were nice reads, but I’m still debating whether to give them my vote.

Brian, I absolutely loved your piece this week. I’d known nothing about AAABA before this (outside of your brief mentions scattered around the web). I didn’t know it was anything significant, let alone a place some D1 schools send their players.

Going to read Matt’s now…

Thanks, Brian.  I really liked yours, too.  You actually write really well when you’re telling stories, despite some criticisms people have when you try to do analytical writing.  I struggle a lot with analytical writing myself, yet also seem to do a good job storytelling too.  I really liked yours this week, and I think it probably gets you a lot of votes. 

I’m somewhat afraid that the clear drop-off in my article from last week to this week might hurt me, but it’s early enough that I’m hoping I get through fine and I worked my butt off on this one, which I think shows (even if obviously Rany Jazayerli did a far better job on some of the same topics, albeit with a longer time frame and a research assistant, a few years back). 

I also thought Kniker’s was awesome again.  Other than mine, which I obviously liked the topic or I wouldn’t have written them, both yours and Kniker’s have consistently been good every week.

One thing that I thought was pretty interesting was that Will Carroll said up front that BP needs to do a better job of cataloging what’s been done, as does the rest of the sabermetric community.  I think there should be some sort of clear database of what research has been done on each topic, and maybe even banks of data that people could use.  I would love if there were spreadsheets available online where people could share data sets they’ve created.  Even if there were a charge for doing it (or some sort of mandatory advertisement requirement for those of us without much money to spend on this stuff), I think that a lot of this “re-inventing the wheel” stuff would be done less.

Brian’s:
http://www.baseballprospectus.com/article.php?articleid=9009

Not at all my cup of tea.  Indeed, Brian simply threw an Euphus pitch to satisfy all those people who are afraid of numbers.  And, it looks like most readers bought it.  It was a typical blog post, and not what I’d want to read in an article.  This article simply proves my point about the competition.

***

Brittany:
http://www.baseballprospectus.com/article.php?articleid=9012

Typical thing you’d read in a newswire report.  BPro is simply not at all for her.

***

Matt S:
http://www.baseballprospectus.com/article.php?articleid=9013

Good research, decent results.  It’s a 1 or 2 article.  I’ll see how it fits compared to the rest of the articles I’ll read right now. 

***

Ken F:
http://www.baseballprospectus.com/article.php?articleid=9015

Decent read.  This is a “1” for me (decent), which makes Brittany a 0, Brian probably a 0, and Matt a 2.

***

Tyler J:
http://www.baseballprospectus.com/article.php?articleid=9014

Not my cup of tea.  I’m sure there’s 8833 stories like this, one for each player in pro ball.

Back in a bit…

Matt K:
http://www.baseballprospectus.com/article.php?articleid=9011

Not my cup of tea.

***

Brian O:
http://www.baseballprospectus.com/article.php?articleid=9010

Decent read. 

***

Tim K:
http://www.baseballprospectus.com/article.php?articleid=9016

Good research, decent read.

***

As a whole, this batch proves the point about the contrived themes.

Updated standings:

6 Brian C 3 3 0
6 Tim K 3 1 2
5 Matt S 1 2 2

3 Jeff E 1 2 x
3 Brian O 1 1 1
3 Ken F 1 1 1
2 Byron L 2 x x

1 Tyler H 0 1 0
0 Brittany 0 0 0
0 Matt K 0 0 0

The lines have been drawn pretty clearly here. 

Do the right thing BPro already, end the contest with 3 or 4 winners.

As for the central database, that’s why this exists:
http://www.tangotiger.net/wiki/

It’s up to people to update it, not just me, Patriot and a few others.

I laughed when I saw that Will Carroll said Brian’s piece this week was the best he’s had so far.

I agree with Sky’s comment: “Anyone a West Wing fan? Remember the episode where Leo McGarry purposely sets the bar low for his VP debate so that when he comes off as competent, he looks even better? That’s how I feel about Brian’s piece this week. He showed a side I think many people assumed he didn’t have, so even though it might be a 7/10, we’re crediting him with an 8/10 or 9/10.”

I didn’t read any of the comments aboout my own article until now, as I was more than a little nervous.

Last week Richard Bergstrom asked why I was clear and concise in comments, but not in the articles. Even back in college, I had trouble being comfortable writing, and I do much better extemporaneously. The more I think about it and try to make it an ‘article’ the tougher it is for the words to come out. I started out writing about Jim Abbott’s game in 1986, but the words just wouldn’t get out of my head and on to the page.

I had previously written a AAABA for Seamheads that was very well received. It’s a topic I know well, and I don’t think I really repeated much from the first article, but I think I’ve used up all my anecdotes.

I took a risk on doing a no-numbers article, but I wanted to show a level of competence in my writing that I didn’t think was apparent the past few weeks, alot of which has to do with the word and time limits. I confess that the week 1 and 2 articles were both done on Thursday nights, leving little tim to self edit and fashion a good closing.

My wife and I both liked it when I filed it, but yesterday and today I started having doubts, and read everyone else’s before I got to mine, even coming here first. However, Kevin Goldstein’s and Christina Kahrl’s comments were exactly what I was trying to convey. So, even though Tango didn’t like it, it accomplished the goal. Next week, it’s likely back to some number crunching, hopefully inside some better prose, as Sky is looking forward to.

Knight - Yawn. An awful lot of effort to tell us that players who are repeat AAA All-Stars don’t tend to be majore league All-Stars. He either needed to go more Pos with it and give us more of a narrative context and some more engaging details, or break out the MLEs and the WARP numbers and do some real analysis.

Cartwright - I read the original article at Seamheads, and I’ve read Brian talking about the AAABA before, so this wasn’t really fresh for me. I understand I’m in the minority on that score.

Funck - Decent content, bad presentation. Unless you’re very, very good at it, that kind of presentational cuteness distracts from your actual content, doesn’t add to it.

Ghiroli - If she has strengths then at some point she needs to start playing to them. Sloppy work, with both glaring factual and language mistakes that really drag the work down; since her supposed strength in this contest is writing and reporting, if she can do neither proficiently then I don’t see what the point is.

Hissey - There are baseball players! They do basebally things! Sometimes they played other sports in high school as well! So what? Why these guys instead of the hundreds of others in the high minors? The article never bothers to establish this. Again - it feels halfway between a fluffy feature piece and analysis and falls well short of both.

Kniker - Fan-freaking-tastic. That said, I would have liked to see a bit more reference to some other work on the topic, like Silver’s stunning analysis of MLB team markets. (Also, win-loss record seems like an obvious regression variable that’s entirely overlooked.)

Oakchunas - A bit of research would poke significant holes in his basic premise that women can in fact play pro baseball; he brings up Ila Borders without acknowelging that she got pretty lit up in a handful of indy leagues:

http://www.baseball-reference.com/bullpen/Ila_Borders

He also fails to draw from the NBA/WNBA experience, despite that fulfilling his ideal of women having the opportunity to play the same sport as the men.

Swartz - Good work, nice presentation, gave enough of the math for you to follow without wading in it. I’m still not persuaded by WARP3 even post-annual and I would have liked to see a nod to Victor Wang’s work, as well as Rany’s.

(Which brings up this comment of Will’s:

“Part of that is the fault of BP and the sabermetric community at large, in really not having one central comparative database of the work out there.”

I think the real problem is that, in a number of ways, BP has removed itself from the sabermetric community at large.)

I think the real problem is that, in a number of ways, BP has removed itself from the sabermetric community at large…

... but a portion of its writers and readers do not know that.

re: Tango’s standing in #213

“The lines have been drawn pretty clearly here. 

Do the right thing BPro already, end the contest with 3 or 4 winners.”

Clearly drawn for you maybe!

I still disagree that the “right” thing is ending the contest early, but as a subscriber I do hope (and expect) that you will be right and 3-4 guys will be hired.

I’d like to see Tango’s top 3 - Kniker, MattS, Brian - continue to write for BP. 

I’d also like to see Ken Funke continue as well and unlike Tango have no problem putting him into that top group.  To me, he is clearly the best writer in the competition and I do think that should matter.  I’ve actually been a bit disappointed that the clearly highly literate readers of this blog have paid so little attention to writing quality in this competition.

Yes, we all want more and better analysis at BP.  We also shouldn’t forget that Bill James had the impact he did in large part because he is a helluva a writer.  And that talent allowed him to be persuasive and a pleasure to read. 

I actually wasn’t keen on this last one by Funke because of the cutesy translations, but some of his others - the good face entry piece in particular - were pleasures to read.

I’m also higher on Matt Knight - nobody should be behind Brittany - because I thought his entry had some potential.  He hasn’t really shown signs of that in the competition and isn’t a top contender, but he’s been readable and decent.

And since I know Matt reads this let me add that he’s been the only one that really significantly changed my first impression.  I didn’t care for Matt’s entry submission, but have loved each of the theme week entries.  My early line was Kniker winning pretty handily, but Matt is going to give him a run for his money.

Maybe it’s just the themes, but I hated the week 2 fantasy stuff (I don’t play), but thought this was a real strong week.  I’m curious to see theme 4.

Colin/#217

Thanks for the reference on looking up Silver’s MLB Market article (sounds like there is also something in Baseball Between the Numbers) so I’ve got some reading to catch up on.  Other thing is that about 4-5 weeks ago there was a pretty good regression article on MLB attendance, but I don’t remember where it was and can’t seem to find it anymore.

Probably the hardest thing about the competition is (at least for me since I have no backlog of work) is getting a topic and trying to put together data and write a coherent article in 3 days.  No real time for a thorough study.

I went with the anecdotal words that Minor League W-L has very little impact on attendance.  If I expanded this into a week or two on-going study, that plus a few other things would be next on the list to try.  I guess I just felt that population, ballpark age, and parent club proximity probably trumped W-L record as likely factors.

Re: attendance, wins, market size, etc.

Voros has a good piece in 2003 or so, as did Keith Woolner, and Derek Zumsteg I think in BPro.com in 2003, maybe 2002.  Also in By the Numbers (at philbirnbaum.com, there was also another good piece, probably in 2007, maybe 2008).  And there have been many more, especially with the impact of the new parks.

philly:

I still disagree that the “right” thing is ending the contest early, but as a subscriber I do hope (and expect) that you will be right and 3-4 guys will be hired.

Are you suggesting that the “right” thing is to crown one winner, and then hire the top 3? 

What exactly was the point therefore of Clay v Reuben in Idol 2, if both got the same recording and marketing deals? At least in Idol 1, Kelly Clarkson got everything, and we’ve been spared any sighting of Justin G.uu.a.rr.i.n.ii. (*) (other than the contrived Kelly/Justin movie, which IMDB.com users unsurprisingly called the worst movie of the year at the time). 

(*) I put in the periods and extra letters on the hope that I will be spared any search result hits from google if someone was actually looking for him.

But, since Idol 2, when it became quite clear that you had at least 2, if not 3 top singers that the muscle would be spread out to the best ones, not just the winner.

Interesting that I should read this comment by Colin

I think the real problem is that, in a number of ways, BP has removed itself from the sabermetric community at large.

on the same day I read this post on Scott Rosenberg’s blog:
http://www.wordyard.com/2009/06/03/once-more-into-pay-wall/

The Journal execs can say, “Hey, we’re just being flexible, it’s a hybrid strategy,” and they’re correct, in a sense. What their strategy fails to take into account is how much traffic and mindshare they have lost from the perception that their articles aren’t a linkable part of the Web.

Has BP’s erection of a pay wall ultimately had the consequence of shutting them out of the main currents of the sabermetric debate?

I don’t think it’s a direct consequence.  I believe it was simply a shift toward the Fantasy market.

Mike, I like the article you linked to. 

I see BPro more like BaseballHQ and other pay-for Fantasy sites, and therefore are part of the niche that is the exception.

"Are you suggesting that the “right” thing is to crown one winner, and then hire the top 3?”

Basically.  Btw, how do you do the actual quote thing?

I think when you start a contest and lay out rules you should follow them to completion.  Although you also need to be flexible, ie hire other qualified people.

I beleive there was a lump sum cash payout to the winner and a weekly slot for some period of time.  Whoever wins should get that.  And then the two or three others who will have hopefully built a following for themselves will be able to negotiate their own deal.

I don’t see a downside.  You’ve mentioned the hazing apsect and the waste of time forcing themes aspect.  None of the contestants have implied that they feel any hazing.  Maybe they can’t be honest about that, but at least one has said exactly the opposite.  It’s true that someone like Brian might be able to spend his time on different topics that he can really push forward, but you never know when constraining someone in certain directions might spark some new interest.

It may be that Matt has always wanted to do draft studies and would have eventually gotten around to doing one, but from Matt’s piece and his comments it sounds like he’s pretty well jazzed about area and has some good feedback to continue.  That’s not necessarily a bad thing.

This doesn’t seem to have any impact on how you’re viewing the contest, but from a marketing/branding point of view this looks like a huge success for BP.  Corporate entities don’t go around pulling the plug on successful marketing campaigns and call that anything but the “wrong” thing to do.

How to quote:

***

I disagree about following to completion.  There should always be something that allows for flexibility.

I disagree about adding constraints to get inspiration.  If this were true, why don’t more editors constrain their writers.  Indeed, quite the opposite is what leads to success.  Writers don’t need shackles to improve.

My basic point stands, that once you are down to the final 3 (Brian, Matt, and Tim), there is no remaining purpose to continue the contest, other than for pure marketing.

Will, Kevin, and Christina, when they will be faced with those final 3 must be asking themselves: “Why the f-ck are we still going.  We surprised ourselves with getting this talent.  Let’s hire them already.”

The rest of BPro authors who are opposed to the contest will be asking them: “Why the f-ck are you still going.  You surprised us with getting this talent.  Let’s hire them already.”

The readership will be asking them at the final 3: “Why the f-ck are you still going.  You surprised us with getting this talent.  Let’s hire them already.”

And the defense is going to be: “Rules are rules?  So, we’ll crown a champion, and then, we’ll hire all three anyway?” Well, that certainly circumvents the spirit of crowning a champion, doesn’t it?

My basic point stands, that once you are down to the final 3 (Brian, Matt, and Tim), there is no remaining purpose to continue the contest, other than for pure marketing.

Actually that’s a bit different than what I thought your basic point has been.  Haven’t you previously said - end this NOW and hire the best 3?

I think that raises big issues of unfairness to the others and would possibly piss off readers/customers who were told that their vote counts.

I don’t like that at all.  I am more comfortable with the idea that the they get to a top 3 and simply acknowledge that all will be hired and stop the contest.

Yes, that is my point.  Near the beginning, I had said:

Put them out of their misery Will… when it gets down to the final 5, hire them all, and announce it.  Split the 1000$ among all 5.
...
The reality is that at the very least, three if not five, of those people will be producing regular or semi-regular freelance work for BPro.  If we accept this as the eventual reality, exactly why are we making Brian and the others do what they are about to do?

Part of that is the fault of BP and the sabermetric community at large, in really not having one central comparative database of the work out there.

If Will means something like a LexusNexus index for sabermetric articles, then yeah, he’s right. Charlie Pavitt was keeping track of articles in the acadmeic realm for a while, but his DB seems to be no more. I don’t know of anyone else who was or is currently doing so.

Week 3 results are in, Brittany is out
http://www.baseballprospectus.com/unfiltered/?p=1307

Week 4 articles are up, everyone is doing a profile of the player of their choice
http://www.baseballprospectus.com/bpidol/

All the pieces were even, decent reads.  No one stood out, and no one sucked.

Honestly, I hate this sort of player profile. It tries to do too much in too little space (with only 2,000 words, every word devoted to which college a guy was going to go to but didn’t costs dearly), and always seems to devolve into presenting data without context. Not to pick on Matt, but his chart of opposite-field home runs - does it really matter? How much? That’s a far more interesting article, isn’t it?

I guess it “sells” well, but I can’t stand any of it; it feels like pouring lead dressed up in numbers and equations. There’s no real meat to any of the analysis.

Colin, right exactly what I was thinking.  You could change the bylines, and you wouldn’t know any better.

As a good example, look at The Fielding Bible II, the player comments section.  The ONLY ones that stood out (for me) were the couple written by Bill James.  All the other ones are interchangeable.

This week proves that BP needs to get rid of the weekly theme. It’s clear that Brian struggles with the topics that can’t be put into a statistical study--but why should that matter? If he gets to write for BP, he’ll never write an article that’s only a narrative with no stats, like last week.

I say BP should let the contestants choose their article topic from here on out.

The time available for this contest is much less than what would be for even a weekly artcile. We get the topics late Monday afternoon, Monday and Tuesday I set aside for research, Wednesday and Thursday for writing.

I picked McCutchen, did the research and number crunching that I could in two days, and really found no compelling statistical story. He’s an average hitter with a very good glove. He actually did hit mroe grounders than average to take advanatge of his speed. Not as good as McLouth, but definitely better than Morgan. If I had another day or two to reflect, I probably should have commented that it might have been better for the Pirates to hold on to McLouth and move him to a corner, with McCutchen’s arrival (always expected at this time, probably a Super2 move) pushing Morgan to the bench.

I do like analyzing and projecting the minor leagues, and I am currently building a statistical and software library to help me in this. Not yet finished, but close, and I dare say better than most have available, especially for free. I wouldn’t mind profiling a minor leaguer each week, gving them 1000-2000 words, while Kevin Goldstein highlights who’s hot. Ehat most often throws me off schedule is deciding on what to write. Of course I love discerning new knowledge, but it’s hard figuring out things no one ever has before every week

Right, I have my own inspiration when it comes to topics and writing.  I don’t want someone else to come in and say “write about this”.  You are not getting the real Brian or Matt when they are put in these contrived themes.

But Bpro will follow the silly letter of the rules on this, and then break the spirit of those rules once the contest is over by hiring whomever they darn feel like hiring.

You don’t think that these themes are telling us which way BPro is looking? They think they need a stathead, are curious if a female or a reporter (or hey! a female reporter) would sell the same way Peroto has. I also think this week’s contest was to see if they could plug the people into the format they already use. Since most of them are as good as what they do anyway, i t hink that was a success. Goldstein’s not stupid and he’s also not a stathead. Watch - in the next couple weeks they’ll have a theme focused on ESPN and we’ll know that this is all planned.

No, I don’t think so at all.  I think BPro Idol has minimal support at BPro (initially anyway), and they’re not going to use anything here as “survey” material.  Their page hits already tell them which author / topic is their most popular feature from their regular writers.

The manager and editor of the site is “minimal support?” If you’re talking about the continued shift of BPro to a subset of ESPN Insider, it’s just another way to marginalise Sheehen who is sure to be somewhere else soon. Kahrl is the only one left unless Davenport is still involved after the book debacle. I think Idol tells us EXACTLY what BPro will be next year.

Well Kara, it’s looking like one of Brian/Matt/Tim are going to win the contest.  All of those guys are statisticians to varying degrees, and hiring one of them would represent a shift away from the ESPN insider type website.

Idol will tell us what the active set of BP subscribers would like, but given that the “contract” only goes through October 2009, there’s no guarantee that the winner stays on

It’s starting to look like the contest is going to last longer than the contract

This week’s articles are up; the theme is “history”.

http://www.baseballprospectus.com/bpidol/

My thoughts:

Brian Cartwright, Stolen Bases: Good contradicting graphs at the end to support the conclusion. Very concise, but the easiest article to follow. Probably the best of the week just for that.

Brian Oakchunas, Babe Ruth: Feels like he spent the whole article either telling us about DTs or showing the steps of how to adjust Ruth’s peripherals and ERA to 2008’s environment. Way too much was spent on DTs, I think. He could have (and should have) spent just one or two paragraphs adjusting Ruth’s peripherals and ERA, and then apply those adjusted numbers to the Ruth/Bonds essay he brought up at the beginning of the article.

Matt Swartz, Championship Expectancy/1991 WS: Easy to follow. I like the idea of CE, but I wasn’t interested in how it applied to the 1991 WS. The best part of the article was this sentence: “For reference, that 29% gain in CE was larger than Kirk Gibson’s 1988 WS HR (+27% CE), Bill Buckner’s 1986 WS error (-20% CE), and Carlton Fisk’s 1975 WS HR (+18% CE).” I would’ve been more interested in the biggest plays in WS (or playoffs) history than the 1991 WS.

Ken Funck, Little Big Man: Ran way too long. I didn’t see the importance of the LBM stat (SLG minus OBP) either. Perhaps the structure could have been improved by looking at player careers instead of decades.

Tim Knicker, Offensive shortstops: Okay, so you picked Win Shares as your production stat. Don’t spend 400 words on why you did! Win Shares may have some flaws, but I don’t care about that if the article has nothing to do with them. As for the article itself, it seemed to jumpy and choppy, and it was hard for me to follow along. I was a little lost after the first graph, when he points out that shortstops weren’t “the defensive golden age as we initially thought”–I didn’t remember what he was referring to (the very first graph, with the spike at the 1970’s fielding value for SS) until I just now was going back over the article. It’s the second best of the week, I guess, but I had to read it twice to fully understand it.

Matthew Knight, W-L records: Echoing what Kevin Goldstein said: Very few BP readers look at pitcher W-L records at all, so why the need to write an article about them? I wasn’t sure how expected W-L records were calculated, but I could still understand the table that showed eW-L records. (I think they were calculated as eW% * average # of decisions for that year.) I had no idea at all what the final table meant, however.

Using my 0-4 star scale (0=not my cup of tea, 1=decent, 2=good, 3=great, 4=I wish I did it)

1:
Matt S
http://www.baseballprospectus.com/article.php?articleid=9090

Brian C
http://www.baseballprospectus.com/article.php?articleid=9091

0:
Ken F
http://www.baseballprospectus.com/article.php?articleid=9093

Matt K
http://www.baseballprospectus.com/article.php?articleid=9094

Tim K
http://www.baseballprospectus.com/article.php?articleid=9089

Brian O
http://www.baseballprospectus.com/article.php?articleid=9092

***

This is still a 3-man race (for me), with Brian C, Matt S, and Tim K.

Brian- Loved the article.  It was simple, concise and the graphs at the end provided an excellent visual representation.  It didn’t seem like a very revealing or ambitious topic (which might be why Tango only gave it a 1), but I enjoyed nonetheless.

The Typo mentioned in post #104 regarding the LF distance for Angels Stadium has been corrected in the latest version of the Ballparks Database.

See:

http://seamheads.com/blog/2009/07/02/the-seamheads-ballparks-database-is-here/

I scanned the new articles from the last 2 weeks really quickly, and nothing jumped out.  If there is something that you guys thought deserves to be highlighted, please do so.

Otherwise, it looks like it’s like a marathon to the finish, and the three guys we all know who are going to be offered a job will be getting a job offer, and the “theme of the week” is simply anti-climactic at this point.

And just like that, things go boom. MattS is out:

http://www.baseballprospectus.com/unfiltered/?p=1330

A couple other thoughts:
1) Why a radio interview? Do they expect these guys to have time for that if the contest is ultimately for a one-article-a-week gig?
2) Interest must be declining, because they’re not seeing this out to a one-on-one finale.

#251

In reply to #1, from Will’s comments, it seems the regular BP writers do frequent radio gigs…

In reply to #2, multiple people posted that BP should hire more than just one finalist, so that might be playing a factor. There were even suggestions in this thread that the challenge should end early and BP should hire many of the finalists. Is there a compelling reason to continue the contest an additional week, a contest that some have called hazing or outright lame, if BP has figured out what their subscribers want and plan to do it? Why prolong the contest if it served its purpose of marketing the site and identified writers that people want to read?

They’ll award the grand prize to one guy, and then hire all 3 finalists.

They put them on their blog:
http://www.baseballprospectus.com/unfiltered/

Or individually (if you are seeing this post weeks from now when they are off the front page):
Tim Kniker’s was my favorite:
http://www.baseballprospectus.com/unfiltered/?p=1336

Didn’t care too much for Ken’s:
http://www.baseballprospectus.com/unfiltered/?p=1337

Brian’s was decent as well:
http://www.baseballprospectus.com/unfiltered/?p=1338

***

I loved Tim’s “swing” / “don’t swing” to the SB breakdown.  I wish I would have thought of that.  The natural extension to that is “don’t swing, called ball or called strike”.  Clearly, a “don’t swing, called strike” is where it’s almost certainly a straight steal.  You could have hit&runs on the called ball ones, or the swings (naturally).

Tim’s second half was already done in The Book.  Chopping that away, his first half of the post was fantastic on its own.  Great stuff.

Brian’s was interesting, but very specific.  Not the kind of timelessness that is my cup of tea.

For me, the winner overall is Tim Kniker (presuming this was the last thing they will be doing), and Brian was just one point behind.  Matt S came in third, but that’s because he didn’t get a chance to contribute this week.

Tim’s analysis wouldn’t include instances where the batter made contact, correct?  So he’s just trying to see if there is a “distraction” effect?  That’s interesting, I guess, but I don’t know how you can separate that from botched hit-and-runs.  And if I was going for a distraction, I’d order a fake bunt and tell the hitter to pull back at the last second, so I’m not automatically giving a strike away.

The finalists also have longer articles, not just the unfiltered posts.

Tango, confirming #256, each person had two entries this week, a quick ‘Unfiltered’ piece from the suggested topics, and then a standard long form article on the subject of the writer’s choice.

Hizouse #256 -

The questioner asked if swinging to protect the runner is a valid tactic, so I broke this down into two parts

1) The benefit of swinging which should be an increased SB rate

2) The cost which is an additional strike

So the first part is to simply say can we see ANY benefit, and while yes it’s going to be tough to tease it out, I think there’s very little evidence that #1 has any measureable benefit.  Therefore, if there is no measureable benefit and only cost, seems pretty logical that “swinging to protect the runner” is a pretty poor tactic.

I got to say, I loved Ken Funck’s article this week, but I wasn’t a fan of any of his prior articles. I’d put Tim or Brian on top, but seeing as how the voters took out Matt last week, I’ve got a feeling Funck wins the competition.

Ken won.

I didn’t even realize that the articles were up.  I’ll check them out now.

Good job by Brian:
http://www.baseballprospectus.com/article.php?articleid=9227

Innovative stuff by Tim:
http://www.baseballprospectus.com/article.php?articleid=9224

Ken starts off with updated data from something from The Book, and then tries to figure out if a paradigm shift is in order.  Probably Ken’s best article.
http://www.baseballprospectus.com/article.php?articleid=9228

I’ll call all these three a wash. 

***

Expect to see Tim and Brian’s work weekly regardless, since, well, it makes sense.

Ken’s article by the way really sidestepped the whole performance issue regarding having pitchers throw on 3 days rest. 

Woolner did show that 3 days rest was better in the time period he studied (70s.80s), while I showed that 4 days rest was best (in the 90s.00s).

In The Book, I proposed keeping the 5-day rotation, but only splitting the tandem approach with the three worst starters (2 innings each).  Ken proposes putting all the pitchers on a tandem approach (4 innings each).

You can make a case either way of course.  It really would depend on how much rest the pitchers need, whether they faced 9 batters or 18 batters.

***

If the M’s out of it on Sept 1, I will beg and plead for Wak to go with an all-relief game at least once a week.  Or even just once.  With the expanded roster, this should be an easy decision.

Tim’s final article is a nice piece.  But it seems to me the strike/ball distinction could be important.  The success rate on non-swings could be in part because some of those pitches are far out of the strike zone, leaving the catcher with a more difficult throw to 2B.  So comparing swings to called strikes might be the most apt comparison.

That said, it seems pretty clear that any gain in SB% is small, and almost certainly outweighed by the cost of the strike.

Guy/263 - pfc gives the opportunity to look at the swing/no swing on the same types of pitches - for example, fastballs in the strike zone