THE BOOK--Playing The Percentages In Baseball

Thursday, May 27, 2010

How lucky has Scott Rolen been with his opportunities to field?

Take the number of balls in play for the league and the team.  See what percentage of BIP the shortstop typically fields, then see how many extra plays Ozzie made.

Right, that’s how WOWY kinda works, and you do end up with “unbelievable” numbers.

I was just looking at Scott Rolen’s career numbers, and whereas the average 3B converts 9% of BIP into outs, Rolen converts 10% of BIP into outs.

Now, if the reason that Rolen is converting so many BIP into outs is because his pitchers, his batters or his park are predisposed to get balls hit to 3B, then Rolen will end up looking better than he is. 

However, his batters, his pitchers and his park have all been virtually league average over his career (as you would kind of expect given a player that has played on as many teams and for as many years as Rolen has).

Now, even given that, perhaps Rolen got lucky and ended up having a disproportionate number of GB/airballs, when he is playing.  That is also not true.

So, we are left with the fact that when there are 4000 BIP per year, the average 3B will make an out 360 times, while Rolen has made an out 400 times, making him +40 plays PER YEAR (+32 runs).

Now, perhaps the reason Rolen is so good is that even given that he’s had average hitters and average pitchers and average parks and average GB/Airball ratios and he’s played so often behind everyone, that even after all that, Rolen is getting a disproportionate share of those ground balls hit to 3B. And that’s why WOWY has Rolen at +170 runs since 2002 while UZR and Dewan has him at +90 to +100 runs.  And that’s why WOWY has Rolen at +300 runs in his career, while Total Zone has him at +150 runs.

So, I ask: what is more believable, that Rolen, given 42,917 balls in play has faced basically an average distribution of plays, such that the league average 3B would have converted 9% of those BIP into outs while Rolen did in fact convert 10%, or that the scorers are so precise in their batted ball locations that an average 3B would have converted 9.5% of those BIP into outs?

The scorers’ marking of the hit locations implies that Rolen’s opportunities are 3.6 standard deviations from the mean.  Now, that, in and of itself, means nothing.  After all, look at enough data, and someone, by luck, will be at 3.6 SD.

Rolen, however, is not an isolated incident.  And so, I submit that the marked locations of the batted balls have enough imprecision in them that a more “earthy” approach, like I’m doing, is warranted to at least regress to.  That even though I’m saying 9% and even though the scorers are implying 9.5%, that maybe it should really be 9.3%.

(The actual numbers I have are this: 42,917 BIP for Rolen, of which he made 4197 outs, or 9.8%.  The average 3B made 8.9% outs.  The average 3B, facing Rolen’s batters made 8.9% outs.  The average 3B, fielding behind Rolen’s pitchers made 9.0%.  The average 3B, playing in Rolen’s parks, made 8.8% outs.  The average 3B, given Rolen’s GB/FB/LD distribution, made 8.9% outs.)