THE BOOK--Playing The Percentages In Baseball

So far, 13 have responded, with 12 saying yes, and the 1 no saying “working for a team”.  A second “no” had pre-responded for the same reason. 

I was really thinking that a few who finished on the low-end would balk, but that hasn’t happened yet.

There’s still about another 20 or so past participants, so we’ll see.  And, no new participants in the queue yet.

Quick question from an observer.  Tango, when you submit your list do you just rank the players based off of their raw points total using your formula?
vr, Xei

In the first year, I used to do points against some positional baseline.  In the second year, last year, I think I didn’t even bother to do that, and simply took total points.  Either way, it didn’t make much of difference.

That’s interesting.  I’d think it would make a sizeable difference as to how you ordered your list based on the distribution of your rankings within each position.  It would be interesting to see how a “smartly” ordered list and a “dumb” ordered list would do against each other with the same set of forecasts.
vr, Xei

Ideally you would “smartly order” which is easy enough, but as the contest is now structured messing with the ordering based on position should have little impact (assuming 2011’s projections are distributed similarly to 2010’s).  I think adjusting the order based on other factors (e.g. upside, injury risk, value gaps at a position) may have more benefit.

From memory, I think it didn’t make much (almost no) difference in 2010 except for catchers.  The OF / 1B / IF / P positions were at pretty much the same baseline.

In 2009 I seem to recall having a lot of 2Bs high on my list.  I do not recall why that was.

Injuries have the biggest impact and it’s tough to deal with them in a drafting format (though it would be nice to have at least one competition that removed players that missed considerable time from the draft pool). There are, I assume, other studies that deal with error on a player-by-player basis. 

I think the next biggest factor, is getting a big season from a player ranked below 250th, and especially below 370 or so, on a draft list.  There was at least one long thread on this last year with a lot of interesting ideas on properly evaluating the impact of these players.  This are of the draft list probably has the most promise for “adjusting”, but this range on the “default” list would contain enough value that a wholesale swap to risk/reward type players is probably a losing proposition.  If I get ambitious with my draft list this is the first place I would look for an edge.

At the moment though, the only thing on my “draft list to-do list” is to make sure Alex Gonzalez is not on it this year.

Yeah, I had a problem with the positional baseline, which is why I finally had the thread I did last week.  When I used average, I got one list, when I used “last player picked”, I got another list.  Then I tried the average of the players from median to last played picked, and I got another list.

Basically, I was not happy with any method I chose.  I definitely soured on “last player picked” because of the “step down” potential.  I had always presumed a nice continuous distribution, but that wasn’t always the case.  It was stupid of me to have presumed that for small samples, which is why I would not go with last player picked.

And, as it turned out, the baselines were so close in some methods that I said: “Why bother?” It’s not like I really needed Marcel to win this.  Marcel was always intended as just a baseline that everyone else should be able to beat.  So, by dialing back my appetite to getting things as perfect as possible and instead going for something fairly simple, I saved myself grief, and provided a decent if not correct baseline.  And, well, we saw what happened.  In one of the contests, Marcel won the darn thing.

We’re at 16 yes, 3 no, so far.  Two of the no as “working for a team”, and one didn’t say the reason.  And still no brand new people wanting to enter.

JEH -

your forecast has done extremely well the last couple of years. any thoughts on getting Fangraphs to include it on their player pages this year, particularly since they won’t be able to include CHONE?

@JEH:  Based on your success in this contest in the past, are you planning on publicizing your projections this year?

@ JEH and @ Mays

yes, like please make it available to Last Player Picked

Mays, hopefully you haven’t abandoned the site?

JEH is posting it at Fangraphs, last time this was discussed.

awesome. sorry if I missed that.

I agreed to provide my projections to Fangraphs to do with what the will this year. 

I am in the dark on the logistics, which isn’t a big deal as I won’t have projections for quite a while. 

I also don’t know which of the numbers that I generate will be of interest to them. I have a lot more confidence in the (dollar) values I generate than in the raw stats so if Fangraphs doesn’t have a place for them I’ll (probably) try and share those files in another way this year.

@JEH:  We’d love to get your projections and run them through our valuation system at RotoChamp, especially to compare our dollar values to the ones you generate.  It would be interesting to see the difference.

@RotoChampMike-

I’ll send the numbers along when I have them, I’m curious as well.

Tango - any idea when we will be seeing the 2011 version of Marcel?

Probably next month…

Just an update: I sent a followup email to past participants for an RSVP.  Just waiting for a few responses.

@JEH

I’ve been eagerly awaiting your projections on Fangraphs, but upon hearing you put more stock in your dollar values, I’d really prefer those.  Would you mind sending them to me?  Thanks!

@20

I’ll be sending FanGraphs projections the end of this week.  Those picks will be completely mechanical, including playing time, and probably not worth a heck of a lot. 

I expect I will send them an update about three weeks from now with manual adjustments.  The vast majority of these updates will affect players that I expect 2011 to be very different from 2010 for one reason or another that I don’t have a good mechanical solution for.

I expect to include $ values for a few common league formats with those projections.  If FanGraphs does not show them (no idea what they include, each projection system they host seems to have it’s unique stat set) I will dump them on the internet somewhere.

@JEH Can you send me an email (mike @ rotochamp.com)?  Thanks, Mike.

@JEH Could you email me please ()?  Thanks!

@ JEH would love to see anything you have

wholecut at gmail.com