Yagyu+

04-17-2010, 12:23 PM

Soto/Hill got mentioned here and there during recent IGTs. Here's an interesting read from ACB (http://www.anothercubsblog.net/2009-cubs-catcher-defense.html). I didn't realize how drastically pitch framing would influence a catcher's value.

Here I'm going to take a quick look at the various components of catcher defense in 2009 for the Cubs' two catchers: Geovany Soto and Koyie Hill. The components I'm going to consider are their fielding and throwing errors, throwing runners out, pitch blocking, and pitch framing.

Fielding/Throwing errors:

These are simply calculated by calculating the number of errors with respect to league average (given playing time) and multiplying the difference by the number of linear weights runs a fielding/throwing error costs. In math, TErrRAA= (TE - (lgTErate * PA)) * -0.48 and FErrRAA = (FE - (lgFErate * PA)) * -0.75. Last season, Koyie Hill saved 0.5 runs with his ability to avoid fielding errors, and was league average in the number of throwing errors he committed. Geovany Soto cost the team 0.8 runs by committing frequent fielding errors, but got half of those runs back (0.4) by avoiding throwing errors. The totals here are +0.5 for Koyie Hill and -0.4 for Geovany Soto.

Throwing Runners Out:

This one uses a similar formula as the last. It compares the caught stealing % to the league-average rate and multiplies the result by the difference in linear weight runs between a CS and a SB: CSRAA = (CS - (lgCSrate) * SBA) * 0.63. Geovany Soto did pretty well here, netting the team 2.2 runs above average with his gun. Koyie Hill did even better, tallying 5.4 runs above average. So far we've covered the "fielding" aspects of catching. Hill's total "fielding contribution" was +5.9 and Soto's was +1.8.

Pitch Blocking:

Again, the number of passed balls and wild pitches is compared to the league average. In this case, the opportunities are balls in the dirt, calculated from pitch F/X data (with HP's help, naturally). Soto cost the team 1.8 runs by not blocking as many pitches in the dirt as the average player, and Hill saved 0.13 runs by blocking pitches in the dirt more average than often. So far Hill is +6.0 and Soto is average at 0.0.

Framing:

Finally, we'll look at pitch framing. Using pitch F/X data, Bill Letson calculated the probability that a pitch would be called a strike or not. Then he gave credit to the catcher for whether or not the pitch was actually called a strike. For example, a pitch on the edge of the strike zone may be called a strike 60% of the time. If a catcher frames one there well enough for it to be called a strike by the ump, the catcher would get credit for 0.4 strikes. If a pitch down the heart of the plate is called a strike 99% of the time, and the pitch is called a ball the pitcher gets "credit" for -0.99 strikes (or 0.99 balls). Corrections are included for things like umpire and the height/handedness of the batter. This is pretty cool stuff, but a bit controversial because the resulting run values Bill calculated were very high. Soto saved the team 5.2 runs in 2009 with his framing ability and Hill cost the team an astounding 16.9 runs. I told you the numbers were high.

Totals:

Basically, it all comes down to framing. That makes sense if you think about it. A catcher will see thousands of called balls/strikes over the course of a season, the majority of which will be around the edge of the strike zone. Any minute ability to frame pitches will make a huge impact, whereas the other stuff happens a handful of times and won't be nearly as important even for huge talent discrepancies.

These numbers tell a surprising story. If you told me that one of the two catchers performed slightly worse at fielding his position (blocking pitches, avoiding errors and throwing out runners) and the other more than made up for it by framing pitches, I would have guessed Soto was better at fielding and Hill was better at framing. But the opposite was true, at least in 2009. What's more, because pitch framing is where catchers derive almost all their value, Soto was a much, much more valuable catcher defensively in 2009. He was worth 5.2 runs above average behind the plate, solely from his ability to frame the plate. Hill, on the other hand, was 10.9 runs below average because he didn't frame the plate well at all. At first I thought there may have been some pitcher effects here, and assumed Hill caught someone more often that was bad at getting calls. But Soto had better framing metrics for the top 5 SP's and the closer; Zambrano, Lilly, Harden, Dempster, Wells, and Marmol all had better success getting calls when Soto was behind the plate. One caveat: Soto was -8.9 FramingRAA/120 G in 2008 and Hill was -0.69 FramingRAA/120 G in 2008 in a miniscule sample size (153 pitches). But this skill was largely repeatable from 2008 to 2009 amongst the entire population of catchers so there's a good chance Hill truly is bad at framing. Soto is likely about average.

If we take these numbers at face value, there was no reason whatsoever for Koyie Hill to start over Geovany Soto. Soto was about 15.1 runs better than Hill on DEFENSE in 2009. Prorated 120 games, Soto was +6.1 runs above average and Hill was -14.3 runs "above" average. That's a difference of 21.4 runs, or over 2 wins in just 120 games! To make up for that, Hill would have had to have posted a wOBA 49 points higher than Soto's. (In case you're wondering, that didn't happen. Soto was 31 points better in wOBA.)

Believe it or not, the best metrics available suggest Geovany Soto was the best defensive catcher on the Cubs in 2009, and it wasn't even close. Throw in that even in a down year Soto was a much better hitter, and there's no question who the best catcher on the team was. Prorated to 120 games, Soto was 69 runs better than Hill in 2009. That's 7 wins! In just 120 games! Hill is bad at hitting, and he's bad at what appears to be the most important thing on defense: framing a pitch. He shouldn't start unless Soto NEEDS a rest.

Here I'm going to take a quick look at the various components of catcher defense in 2009 for the Cubs' two catchers: Geovany Soto and Koyie Hill. The components I'm going to consider are their fielding and throwing errors, throwing runners out, pitch blocking, and pitch framing.

Fielding/Throwing errors:

These are simply calculated by calculating the number of errors with respect to league average (given playing time) and multiplying the difference by the number of linear weights runs a fielding/throwing error costs. In math, TErrRAA= (TE - (lgTErate * PA)) * -0.48 and FErrRAA = (FE - (lgFErate * PA)) * -0.75. Last season, Koyie Hill saved 0.5 runs with his ability to avoid fielding errors, and was league average in the number of throwing errors he committed. Geovany Soto cost the team 0.8 runs by committing frequent fielding errors, but got half of those runs back (0.4) by avoiding throwing errors. The totals here are +0.5 for Koyie Hill and -0.4 for Geovany Soto.

Throwing Runners Out:

This one uses a similar formula as the last. It compares the caught stealing % to the league-average rate and multiplies the result by the difference in linear weight runs between a CS and a SB: CSRAA = (CS - (lgCSrate) * SBA) * 0.63. Geovany Soto did pretty well here, netting the team 2.2 runs above average with his gun. Koyie Hill did even better, tallying 5.4 runs above average. So far we've covered the "fielding" aspects of catching. Hill's total "fielding contribution" was +5.9 and Soto's was +1.8.

Pitch Blocking:

Again, the number of passed balls and wild pitches is compared to the league average. In this case, the opportunities are balls in the dirt, calculated from pitch F/X data (with HP's help, naturally). Soto cost the team 1.8 runs by not blocking as many pitches in the dirt as the average player, and Hill saved 0.13 runs by blocking pitches in the dirt more average than often. So far Hill is +6.0 and Soto is average at 0.0.

Framing:

Finally, we'll look at pitch framing. Using pitch F/X data, Bill Letson calculated the probability that a pitch would be called a strike or not. Then he gave credit to the catcher for whether or not the pitch was actually called a strike. For example, a pitch on the edge of the strike zone may be called a strike 60% of the time. If a catcher frames one there well enough for it to be called a strike by the ump, the catcher would get credit for 0.4 strikes. If a pitch down the heart of the plate is called a strike 99% of the time, and the pitch is called a ball the pitcher gets "credit" for -0.99 strikes (or 0.99 balls). Corrections are included for things like umpire and the height/handedness of the batter. This is pretty cool stuff, but a bit controversial because the resulting run values Bill calculated were very high. Soto saved the team 5.2 runs in 2009 with his framing ability and Hill cost the team an astounding 16.9 runs. I told you the numbers were high.

Totals:

Basically, it all comes down to framing. That makes sense if you think about it. A catcher will see thousands of called balls/strikes over the course of a season, the majority of which will be around the edge of the strike zone. Any minute ability to frame pitches will make a huge impact, whereas the other stuff happens a handful of times and won't be nearly as important even for huge talent discrepancies.

These numbers tell a surprising story. If you told me that one of the two catchers performed slightly worse at fielding his position (blocking pitches, avoiding errors and throwing out runners) and the other more than made up for it by framing pitches, I would have guessed Soto was better at fielding and Hill was better at framing. But the opposite was true, at least in 2009. What's more, because pitch framing is where catchers derive almost all their value, Soto was a much, much more valuable catcher defensively in 2009. He was worth 5.2 runs above average behind the plate, solely from his ability to frame the plate. Hill, on the other hand, was 10.9 runs below average because he didn't frame the plate well at all. At first I thought there may have been some pitcher effects here, and assumed Hill caught someone more often that was bad at getting calls. But Soto had better framing metrics for the top 5 SP's and the closer; Zambrano, Lilly, Harden, Dempster, Wells, and Marmol all had better success getting calls when Soto was behind the plate. One caveat: Soto was -8.9 FramingRAA/120 G in 2008 and Hill was -0.69 FramingRAA/120 G in 2008 in a miniscule sample size (153 pitches). But this skill was largely repeatable from 2008 to 2009 amongst the entire population of catchers so there's a good chance Hill truly is bad at framing. Soto is likely about average.

If we take these numbers at face value, there was no reason whatsoever for Koyie Hill to start over Geovany Soto. Soto was about 15.1 runs better than Hill on DEFENSE in 2009. Prorated 120 games, Soto was +6.1 runs above average and Hill was -14.3 runs "above" average. That's a difference of 21.4 runs, or over 2 wins in just 120 games! To make up for that, Hill would have had to have posted a wOBA 49 points higher than Soto's. (In case you're wondering, that didn't happen. Soto was 31 points better in wOBA.)

Believe it or not, the best metrics available suggest Geovany Soto was the best defensive catcher on the Cubs in 2009, and it wasn't even close. Throw in that even in a down year Soto was a much better hitter, and there's no question who the best catcher on the team was. Prorated to 120 games, Soto was 69 runs better than Hill in 2009. That's 7 wins! In just 120 games! Hill is bad at hitting, and he's bad at what appears to be the most important thing on defense: framing a pitch. He shouldn't start unless Soto NEEDS a rest.