THE BOOK--Playing The Percentages In Baseball

I was going to respond to the comments on the Star Tribune article that Neyer linked to about their assertion that Hunter > Cameron (he’s not, obviously), but decided it just wasn’t worth it.

I’m surprised that people even in Minnesota forget how good Corie Koskie was.

Elk: the way it works is this:

Assigned Fielding
= Actual Fielding / 3
+ Flashy Fielding * 3

I presume Koskie was not a flashy fielder, and so all of his fielding value was probably set to “good, but”.

I went through this on a BTF thread, comparing Morneau’s last 4 years with Koskie’s first 4 years of the last decade.

http://www.baseballthinkfactory.org/files/newsstand/discussion/rand_is_joe_mauer_the_40th_best_baseball_player_of_the_decade/

Looks like using UZR instead of TZ makes the difference even greater.  If anything I’m giving Morneau every benefit of the doubt here.  If I’m wrong I think you need to make one of 2 cases:  Koskie was not a plus defender, or we should use a much smaller position adjustment between 1st and 3rd.

As for Joe Mauer, the question really comes down to is WAR the appropriate answer to “best of the decade”.  Are 4 years of 6 WAR more impressive than 8 years of 3 WAR each?  If so, then you just need to re-baseline.

The Mauer problem is a great illustration of why I hate the arbitary ten-year period being taken seriously.  Mauer’s best years will probably be something like 2006-2015 (at least as opposed to 2000-2009 or 2010-2019).  Hanley Ramirez too.  Meanwhile, Albert Pujols career starts in 2001 as a MVP candidate and thus fits nicely into a calender decade. 

So what?  Mark Grace gets to be a trivia answer because he had the most hits from 1990-1999, but who has any idea that Tony Gwynn led from 1989-1998 and Rafael Palmeiro led from 1991-2000?  (Not that I think hits in a 10 year period is particularly interesting, but you get the idea).

I remember Koskie as a great fielder who made lots of “web gems” way before I had heard of UZR, or ZR for that matter.  I remember complaining when he didn’t win the gold glove.  I think he was underrated offensively in Minnesota - I remember a lot of people complaining about him for some reason.  I think his approach (I feel like he took more than his share of called 3rd strikes) bothered people. 

So I think its partly offensive, but mostly not understanding position adjustments and fielding runs.  It’s not that we didn’t know he was a really good fielder, we just didn’t know how much value that had.

I don’t have any problem with the “best of the decade” lists.  They are what they are and they answer perfectly reasonable questions. No one is claiming (that I am aware of) that a person who IS or is NOT the best in a decade is necessarily a great player or not in other decades.  Most people realize that a player could have been great in parts of two decades and another very good player could just have happened to have his best years within a certain decade.  Those things are obvious. So what?

The Mauer problem is a great illustration of why I hate the arbitary ten-year period being taken seriously.

The “most people” you refer to aren’t the ones who are taking it seriously.  Nor is Rob.

Taking it seriously for what? It is what it is - seriously.  The most productive players during the time period x-y.  What’s to not be taken seriously?  The question is quite clear. There is no ambiguity.  And the “lists” answers the question precisely. Who were the “best” (however you want to define that) players of the decade, whatever time period the decade encompasses. 

What exactly is not to be taken seriously?  Just because you don’t like that question doesn’t mean it isn’t to be taken seriously.  When you say something “shouldn’t be taken seriously” that almost almost always means that something is wrong, as in the answer did not properly address the question.

How is, “Who were the best players from 1990-1999 any different from, “Who were the best players this year, or last year, or whenever?”

Is the question, “Who was the best player in 2009?” not to be taken seriously as well?

It would have been better had I said “Treated as having more meaning than any other ten-year period”, but I thought the rest of my initial post made that point.  Oh well.

What I mean is that the question “Who was the best player from 1990-1999?” is elevated (by some people) to having some greater meaning--like using it as a lynchpin for someone’s HOF case (Jack Morris) or picking “He won batting titles in three decades” as one of the most telling things you could say about George Brett’s career (it’s on his HOF plaque).

The question “Who was the best player from 1990-1999?” is a perfectly legitimate one.  But some people treat it as being infinitely more important than “Who was the best player from 1994-2003?”

In the context of discussing Koskie’s fielding, it’s interesting to note that his career ended on a fielding play, and a great one at that, as Koskie fell backwards while catching a pop fly, the impact with the ground causing his glove hand to fly up and flip the ball back into the air, whereupon Bill Hall snatched it for the out. I still see that replay occasionally and wonder how many people remember just what the true impact of that play was.

Patriot, fair enough…

I have recently discovered something that many of you might have already known, but I find it interesting in my quest to “enlighten” people when it comes to baseball player evaluation.

A lot of people are on board with offensive statistics, and while they don’t all “buy into” some of the defensive metrics, they at least get what they’re about. So, at the very least, I can get on a fairly equal ground when it comes to that (even if they hate the metrics, they understand the idea of defensive runs/wins, which I think is what matters most). Many lay people can look at the batting stats and trust the math. And they can look at the defensive numbers at least hypothetically (IOW, they might not accept that so-and-so is 20 runs above average defensively, but they’re fine with the concept).

Where people tend to have the biggest problem is in positional value. I’m not sure why, since most people would agree finding a good C/SS/CF is a lot harder than finding a 1B/DH, but the hurdle is there.

There was recently a discussion on a local sports radio show, one with a host who is somewhat saber-minded. They were looking at, I think, 2009 WAR and they were all struggling with the idea that in 2009, Nyjer Morgan was worth more wins than Ryan Howard (I believe the numbers were so close they were negligible anyway). I wrote to one of the guys on the show trying to explain that - assuming we trust the defensive values - the reason Morgan is worth more is because of position. I never heard back, so I have no idea if my explanation worked.

The same thing happens with Koskie and Morneau, but it’s even worse because many people still think of 3B as “1B on the other side of the field.” If you asked the average person to do a defensive spectrum, I wouldn’t be surprised if 3B and 1B were at the bottom. So the reason, I think, that Koskie couldn’t possibly be better than Morneau, to most, is that the positions are about equal, and his defensive superiority doesn’t exceed his offensive inferiority.

So for those of us genuinely interested in teaching all this to people, shouldn’t we perhaps start with how important the position played is in determining player value? I feel like it’s often assumed as a given around here, and perhaps rightfully so HERE, but I find that it’s assumed pretty much everywhere, too.

(As a sidenote: I’m also trying to apply a lot of this stuff to a co-ed softball team: The differences are interesting, since catcher is the least important position, first base is important because simply catching the ball every time it’s thrown isn’t something everybody can do, etc. Took me a full season just to convince the team that defense matters though, so it’ll be a while before they understand the other stuff).

Instead of positional adjustments explained defensively, I calculate each batter’s runs above the replacement level offense at his position, which is then proportional to his plate apearances.

The average 1b hits much better than the avg cf. It’s easy to find a 1b (34 of them) with a .354 wOBA, while there’s only 10 centerfielders who hit that well.

There are 120 1b with a wOBA greater than .328, only 41 CF.

The 1b is much easier to replace offensively.

This is my comparison of Howard vs Morgan

Name   RAR Rank  PA Pos wOBA  Avg BRAA Repl BRAR FRAA SBRAA
Howard  46  34  690  1B .389 .354 27.9 .331 46.1 -0.6 -0.1 
Morgan  25 139  504  CF .328 .318 -5.8 .300 10.3 12.6  2.1

I project Ryan Howard to have 690 PAs with a wOBA of .389, Nyjer Morgan 504 PAs with a wOBA of .318, greatly lower.

Howard plays 1b, where MLB average wOBA is .354, replacement level is .331 (93.5% of avg - this is from Tango, but I don’t have the link handy). Howard’s .389 wOBA over 690 PA will generate 28 more runs than an avg 1b, 46 more runs than a replacement 1b. I project Howard -0.6 Fielding Runs compared to the avg 1b, and -0.1 SB runs above average, good for a total runs above replacement of 46, which ranks 34th in my 2010 projections.

Morgan plays CF, where MLB average wOBA is .328 and replacement level is .300 (91.5% of avg). Morgan is thus -5.8 runs below an avg CF over 504 PAs, +10 runs above replacement. Add +12.6 fielding runs above average and +2.1 SB runs above average for a total of 25 runs above replacement, which ranks 139th.

I project the Phillies top position players (guys they owned at the end of the season) as Utley 80, Howard 46, Werth 42, Rollins 38, Ibanez 33, Victorino 28, Michael Taylor 19, Ruiz 14, Travis D’Arnaud 13, Neil Sellers 13, Domonic Brown 13.

My top position players for the Nats are Zimmerman 47, Willingham 34, Dunn 33, Morgan 25, Danny Espinosa 20, Willie Harris 17, Guzman 17, Dukes 16, Ian Desmond 13, Justin Maxwell 12.

Overall, Pujols 105, Hanley 97, Mauer 90, ARod 84, Utley 80.

I have very few position players above 30 still in the minors, and they are top prospects - Jesus Montero 48, Mike Stanton 37, Jason Heyward 33 are some.

From 10 to 20 are maybe half MLB, half minor leaguers.

Howard is 7th among 1b at 46 (behind Pujols, Teixeira, Gonzalez, Cabrera, Fielder, Berkman), while Morgan is 28th among CF at 25. That’s enough for Morgan to be an everyday player in MLB, but he better not drop much.

Brian, if the “offensive positional adjustments” is used only as an illustration, then fine.  If you use it in practice, you are wrong, as we’ve discussed in this blog several times.  In one sentence, what do you do when the offensive level is higher for CF than for 1B in MLB (yes, it’s happened), or higher for SS than 2B in college?

Secondly, I would not bring in replacement level at all in the discussion being asked about.  Calling Ryan Howard a certain number of runs above the average for his position on offense and a certain number of runs below average on defense is perfectly fine as an initial starting point for discussion.  That’s where it should end, because it’ll take 20 minutes for the person upon first hearing it that way to try to wrap his head around it.

Corey Koskie was not better than Justin Morneau. Silly claims like that will keep sabermetrics on the periphery. What the field needs are practitioners who not only understand the numbers, but also understand the game. That is not the case here.

Thank you for taking the opposite side.  Now, can you make your best argument to support your position?

Here the data to use:

http://bbref.com/pi/shareit/yAcTZ

That’s all 89 players with 3000 to 4000 PA from 2000-09/

Morneau’s line: .280 .350 .501
Koskie’s line: .273 .367 .459

A 17 point advantage for Koskie on OBP and a 42 point advantage for Morneau on SLG.  That’s close to a wash.  Add in the fact that Koskie was a much better fielder at a more premium position, and the argument for Morneau will boil down to using the word “RBI” at least 12 times.

By the way “on the periphery” of what exactly?  Of mainstream fan acceptance?  I hope that would be the case.  If sabermetrics was no longer on the periphery, it would no longer be on the cutting edge.

For whatever it’s worth, Win Shares has Koskie and Morneau with the identical totals for the decade 2000-2009: 107 Win Shares each. Morneau totaled 90 in his best four Win Shares years (2006-2009), Koskie totaled 78 in his best four Win Shares years (2000-2003).

Don’t feed the troll Tango.