THE BOOK--Playing The Percentages In Baseball

kind of like the survivor method, where you have every player’s season available and you vote off the least desirable season until the last remaining player’s season is your mvp, just in reverse.

Let’s do it.

Tom, can you start a poll with a “6-pack of MVP candidates” and then we vote them off one at a time until there is one left?

I’m down.

Be willing to do the same thing with AL/NL CYA, NL/AL ROY, etc.

As far as MVP, Am I the only weary of using Fangraph WAR because UZR has such a high margin of error? Jacoby Ellsbury’s WAR is out of this world, but I’m not a fan of 1 year UZR because of the luck and randomness involved in it. Plus, I would much rather use WPA than UZR for MVP voting because that is real contribution we can measure, not theoretical wins vs. a theoretical replacement level player.

For example, WAR would tell you that Brandon Phillips was 7th most valuable player in the league with a 6 WAR. WPA tells you that Tyler Clippard is the 7th most player with a 5.01 WAR (leading all pitchers). But WPA has Phillips at 1.47 above average and Clippard has a 1.2 WAR. I would much rather take Clippard’s contributions because they were more meaningful.

While Clippard only faced 329 PA to Phillips 674 PA, I think holding opposing batters a .535 OPS in 329 PA is more impressive than 674 PA of .810 OPS especially when you consider the high context of Clippard’s appearance.

btw, why is the Fangraphs and B-R WPA different? Aren’t they measuring the same thing?

I think using a salary cap for the draft (average league payroll?) will give a real perspective about what is the real value. Miguel Cabrera salary (20 millions) it´s good enough to pay Pedroia (5.5 MM, 8.0 WAR), Ellsbury (2.4 MM 9.9 WAR), Bautista (8 MM, 8.3 WAR) and Shields (4.25 MM 4.9 WAR) for example, 3 top 5 in the MVP´s count and 1 top 5 in my CY´s count, not bad.
The problem with my fellow venezuelan Miguel Cabrera is the way his salary hurts Detroit regarding his real value.

Tom, I like the method - only one nit: when I used to run an Earl Weaver baseball league in college (long ago!), the identity of the first pick was very often driven by the gap between #1 and #2 at whatever position had the largest gap.  So, if the top 2B was way better than the #2 2B, he’d get picked first, ahead of (let’s say) the top 1B, who was the best in the league, but just happened to be only a little better than the #2 1B.

Does that make sense?

Anyway, I can easily set that little bit aside and get on board with your idea, but wanted to mention it…

So, strato-o-matic is running the sim based on their individual stats not based on “true ability”?  I think this adds some clarity but still leaves some questions on answered - which stats is this sim using and why?  Will the best pitcher be the one with the lowest RA, lowest ERA or lowest component ERA?  I’m guessing that if Granderson’s R and RBI don’t help him but his actual OBP and SLG are used, that we’re looking at component ERA.  If I get to use my own best judgement about how much of Granderson’s RBI total was luck, why don’t I get to choose how much of his OBP was luck?  Do I just end up estimating his true ability?

If you want to include clutch (the timing of his performance), then that’s your choice.  So, Grandy’s RBI are in play as a byproduct of whatever his performance was with men on base.

At an extreme, imagine that all of CC’s hits and walks allowed were with bases empty.  And with a man on base, his OBP was .000.  So, you can count that for his strat card.  (His only runs allowed will be his HR allowed.)

***

As for WPA being different: the only difference should be the park adjustment, however David and Sean do those.  There MAY be a league adjustment, but I don’t know.

But, when I evaluate his performance a hit has to count as a hit?  What if I think someone else hit the ball harder but had fewer balls drop in or someone else took better swings but faced tougher pitches, do I get to adjust the basic stat line too?

5,

That would only matter if you were picking a full team.  To me, this is really more like you only get one player. 

Or, if you like the idea that you’re picking a full team, think of it like you’re pick a full team in a 30 team league.  So then your nit won’t matter much, since your first pick will have basically no effect on your second pick, so you’ll just pick the best player.

J: I guess it’s whatever you can reasonably argue.  If a player constantly gets HR robbed by Bourjos and Guti and Jackson, how do you want to count that?

For MVP purposes, I’m thinking we’re going to count that as if he did not hit any HR.

But, you can argue whatever you can.

Sorry to be difficult.  I just have trouble envisioning a logically consistent approach that isn’t either WPA-like or, alternatively, just a true talent estimate.

Why do the “WPA = MVP!” crowd always forget positional expectation?

J.Cross - I really don’t think we should credit a guy for bad luck when giving an award based on what happened.

As far as I’m concerned, MVP isn’t for true talent, but for on-field results. The award isn’t “most valuable going forward” after all. Consider this: One player gets all bloop singles, another makes all hard line drive outs (or even has 550 ABs of robbed home runs). While the unlucky guy is the better hitter, the lucky guy did a lot more to actually help his team that season.

Here’s the play you have to decide in terms of “crediting” a player’s performance:

http://mlb.mlb.com/video/play.jsp?content_id=7143405&topic_id=&c_id=mlb&tcid=vpp_copy_7143405&v=3

That’s Luis Gonzalez’ World Series winning hit against MAriano Rivera that scored Jay Bell.

By my count, that ball was in the air about 3 seconds, and it travelled 140 feet.  The only reason the fielders were playing in was because you had a runner on 3B.

So, the runner on 3B gets “credit” for drawing the infield in.  Rivera gets “credit” for giving up the weakest 3-second non-out in the history of baseball, and Gonzalez gets “credit” for getting the ball to land in no-man’s land.

If let’s say the MVP considerations, prior to that play, was a three-way tie between Gonzo, Rivera, and Jay Bell, then after this play, who gets your vote?

#12- I don’t see where positional makes a big impact on WPA. Bautista led baseball with his 7.86 WPA. I don’t care about positions because he made a real measurable impact. Positional adjustments are just theoretical hypothetical with huge margins of error. Who is to say that Jose Bautista wouldn’t have provided the same value at SS with his bat? Facts are Bautista provided the most “real” value (IE: WPA) than any other player, not hypothetical value.

I like that thought experiment.

I’d argue that Louis is the MVP, because of the combination of intention and action.

Mariano intended to get the out. He did not succeed, however unlucky the result may have been.

Jay Bell did bring the infield in, but that was not his personal intention, merely a natural result of his being on-base (at least with respect to where Louis’ eventual hit lands with respect to the defense, unless Jay is clairvoyant).

Louis, however, both intended to get a hit, and did produce a hit. The components of luck involved should not be counted against him.

Obviously, this is not a measure of likely future success, or a prediction of outcome should the situation between Rivera, Bell, and Gonzalez be repeated, but only a recognition of credit based on the combination of what each party wanted to do, and what each party actually accomplished.

Ok, so what if Gonzalez hit the ball straight to Jeter on one hop, and then Jeter made a bad throw to the catcher, allowing Jay Bell to score.

Supposing Bell, Rivera, and Gonzo were all tied for MVP heading into that play, who gets your MVP?

pm/15: the WPA you quote is only for offense.  You are not accounting for fielding WPA.

Interesting concept, but a few things I have apprehensions about.

1. Age of player.  I’m drafting a younger player in most cases.  So this wouldn’t be fair to an older player who edged out a few other players.

2. MVP should be for “one” season (this season!).  If I am drafting one player, I am likely to consider multiple seasons to determine true talent level.

3. Might be semantics but this should probably be a different award like “Best Player” instead of 2011 MVP etc…

You are only drafting for the one season, and this is after knowing after-the-fact his performance line, overall, and in various split scenarios.

So, I don’t see why age would matter at all.

Tango, so why not add WPA and defensive contribution relative to average, and that gets you the MVP.

How much do you factor in WPA in your MVP vote? This obviously isn’t about true talent, but the guy who contributed the most value in a season.

I think it’s reasonable enough to do as you say (thought obviously you’d have to add the positional adjustment… an average fielder at 1B is not the same thing as an average fielder at SS).

I have done this for a while. Just take the Fangraphs breakout of WAR into a spreadsheet, replace RAA with WPA x 9.5 (or 10), and then add it up. You dont have WPA for fielding or baserunning, but you have positional adjustments.

#20. Sorry I didn’t see that in the blockquote anywhere.  Given that it is a draft for only the next year and then the players are re-drafted again that changes things.  But I still think it is highly flawed.  Let’s say you have one dominant player, heads above the rest of the league.  Let’s give this player the name of Barry Bonds for namesake.  Let’s say Barry dominates for half the year, then either gets hurt or fails a drug test and misses the rest of the season.  The injury is not bad enough to be debilitating.  He likely goes number one in the re-draft but didn’t play enough to deserve the MVP.  Age won’t matter, true.  MVP should be based off of the previous season and not weighted on the past 2 or 3 which is what you are doing when evaluating players for the re-draft.  Interesting but flawed.

JD - If we’re crediting a guy for bloop singles b/c after all, they fell in, why not credit pitchers for wins and losses and hitters for R and RBI?  I feel like the two extremes are WPA-like (yes, position adjusted with defense added in and what not) and true talent.  Anything in between and you’re choosing to include some kinds of “luck” and not others (and no, I can’t really define luck here).

Tango/14 - I like the hypothetical!  I think an average runner draws the infield in so I don’t give the runner any credit.  I guess I’d have to give credit to Louis and give him the MVP although I’d admit that’s at least a little ridiculous.

Ok, now I am confused.  Is this mythical draft for the 2012 season or are we going back in time and re-drafting to play the 2011 season over?  This makes a big difference and is not spelled out in the blockquote or post.

Xei: I think you should read Poz’s post if you need more context.

Tom, I like the method - only one nit: when I used to run an Earl Weaver baseball league in college (long ago!), the identity of the first pick was very often driven by the gap between #1 and #2 at whatever position had the largest gap.  So, if the top 2B was way better than the #2 2B, he’d get picked first, ahead of (let’s say) the top 1B, who was the best in the league, but just happened to be only a little better than the #2 1B.

Does that make sense?

It makes perfect sense to me, at least in the example you gave. I think it is at least arguable the the 2B was the most valuable player, because his level of talent was much harder to obtain.

The voters at times have recognized that to a degree, awarding the MVP to catchers and shortstops who didn’t have the best overall offensive production.