THE BOOK--Playing The Percentages In Baseball

Tuesday, December 29, 2009

Mike Silva Chronicles - Part 2: WAR

WAR makes assumptions about the value of defensive positions, assumes a certain “replacement level”, and assigns a “win share” to players as if they perform in a vacuum. First, where did you come to your conclusions regarding the value of defensive positions?

The value of the fielding positions was based on how players performed at multiple positions.  For example, if Brian Giles plays LF or RF, and he has the same stats relative to the average at those positions, then I conclude that LF and RF are equals.  If Ichiro plays CF, and I find his stats are 10 runs worse than average at CF than in RF, then I conclude that the competition level in CF is higher than in RF.

When it comes to players who play multiple positions, they are in abundance in the outfield.  And there’s limited bias going on.  And that relationship shows that LF and RF are about equal, and the average CF is about 10 runs better than the average corner outfielder.  There’s really no surprise here.

I apply the same process for SS, 2B, and 3B.  Here’s it’s a little dicier, because the players who play at multiple positions are both less than the outfielders, and more selective in who actually plays the multiple positions.  Understanding that there is some level of uncertainty here, my current conclusion is that SS is about 5 runs ahead of 2B and 3B, with 2B and 3B being equals.  Like I said, if someone wants to dispute it, I don’t have much disagreement if you want to move things up or down 1 or 2 runs.  Maybe 3.  Anything more, and it’s stretching it.

Then, looking at the IF/OF translation is a little tougher, for two reasons.  One, very few true outfielders become infielders.  Almost all the moving is from career infielders becoming outfielders.  So, when you look at how the players perform at IF/OF, you have a bias, because there is a certain “familiarity” factor that has to be accounted for.  Secondly, no LH will play in the infield (2B, SS, 3B).  So, that reduces the pool of players.  So, I try to reason it out, and I had a few thread that showed how I went through it all.  I also relied on how they did as hitters, reasoning that IF/OF are distinct enough that the hitting talents should offset the fielding shortcomings (and vice versa).  Not totally of course, but it served as a useful guide.

I really relied on this for catcher v everyone else, since catcher is a pretty unique position.  So, I pretty much made the catcher’s fielding contribution the opposite of their hitting contribution.  And 1B falls on the end of the spectrum.  For DH, I reasoned that a DH is equal to a bad fielding 1B.  So, the spectrum becomes this (in wins form, which is runs/10):
+1.25 C
+0.75 SS
+0.25 2B/3B/CF
-0.75 LF/RF
-1.25 1B
-2.25 DH

Second what exactly is replacement level? Where did you come to that baseline?

All it is is the contributions of players that are not part of the 25-man roster, basically.  You can also look at it as the best (non-prospect) AAA players.  There’s been many studies on this issue, and the consensus is very close to being 2 wins below average.  MGL for example uses 18 runs below average per 150 games.  Keith Woolner uses 20 runs per 162 games.  I use 2.25 wins below average per 162 games.

So, an average player who plays 108 games would have a WAR of +1.5.

Finally, assigning a win share to a player assumes said player would perform the same on another team. Knowing the factors such as protection in lineup, ballpark, etc. makes me less likely to understand the value of this stat in player evaluation.

We presume that a HR hit is a HR hit, for the purposes of assigning value.  It does account for the ballpark, but it does not account for “protection”.  Protection is real, and it does exist.  But, in an overall value sense, it has almost no impact.  That is, while protection will affect how a pitcher pitches to a batter, and how many K, HR, BB a batter gets, these things end up balancing each other out to such an extent that, overall, it has no impact.  I know, hard to believe, but that’s why we research these things, and present the evidence.  Andy wrote it up in The Book, and as luck would have it, we excerpted it for all to see:
http://www.hardballtimes.com/main/article/pitching-around-batters