I came across this because I’m working on a simulation game with players rated based on their stats. In my attempt to rate each catcher’s throwing arm, I decided to use their caught stealing percentage, but I didn’t want their battery mates to have an affect on the ratings. That is why I decided to use the WOWY method (with or without you) which compares the caught stealing percentages of each battery mate with and without a particular catcher.
After finishing my queries, I like to check some of the All-Time greats to see where they stand. This is when I noticed that Ivan Rodriguez’s WOWY rating wasn’t as amazing as I expected.
Pudge’s peak in throwing out base runners lasted from about 1991 – 2001 where he consistently threw out 20+% more base runners than league average. During that same time, the WOWY method has him 10-15% better than average. That’s still a fantastic rate, but it shows that his overall numbers may be a little deceiving.
The most obvious reason for the disparity would be that his battery mates were better than average, and this seems to be the case with Pudge. Here is the data from ’91-’01:
Year
WOWY
CS
Att
CS%
Pitchers
League
1991
-12.3%
89
176
51%
38%
33%
1992
-5.4%
140
290
48%
43%
33%
1993
-1.2%
130
278
47%
46%
34%
1994
-5.0%
108
241
45%
40%
32%
1995
-11.1%
102
215
47%
36%
31%
1996
-11.0%
127
242
52%
41%
30%
1997
-12.9%
138
253
55%
42%
32%
1998
-13.0%
134
239
56%
43%
32%
1999
-15.0%
104
191
54%
39%
31%
2000
-13.7%
93
165
56%
43%
31%
2001
-10.2%
65
129
50%
40%
31%
Take the 1998 season for example. From ’97-’99 (I use a 3-year average to increase sample size and smooth out seasons), Pudge threw out 134 of 239 runners, 56%. League average was 31.5% during that time. The pitchers he caught most frequently (Rick Helling, John Burkett, Darren Oliver, Bobby Witt, and Aaron Sele) combined for a 43.8% caught stealing rate with catchers other than Pudge. So while Pudge was 25% better than league average, he was just 12% better using the WOWY method.
While the pitchers changed over the course of his peak seasons, the story seemed to stay the same. He caught a bunch of pitchers who had better than average caught stealing percentages.
A couple of notes:
-I count double steals as just 1 steal since the catcher only has the opportunity to catch 1 runner. This is why there may be a slight difference in SB/CS numbers.
-One problem with WOWY in this case is the “WY” (without you) in that the other catchers may also be above or below league average. But this is why I use a 3-year average to increase the sample of battery mates. One could take an extra step and look at the other catchers in the “WY” data to see if/how they skew the numbers.
-This also works for pitchers, and a similar comparison is Chris Carpenter. From ’05-’07, he had a 75% CS%, about 45% better than league average. But his primary catcher was Yadier Molina. Using the WOWY method, Carpenter’s CS% was 28% better than average, rather than 45%.
Much like Johnny Bench before him, Pudge’s statistics don’t REALLY do him justice when it comes to CS, though. One of the biggest factors, for both of these catchers, is how much SB ATTEMPTS were squelched with them behind the plate. I picked a random year – 1999, when I-Rod was with the Rangers. In that year, there were 1460 SB allowed in the AL, with 715 CS by AL catchers (and yes, I realize this was after the advent of interleague play; we’ll just ignore it for now). That’s, roughly, 104 SB and 51 CS per defensive team. Texas had 52 CS – more or less EXACTLY the league average. But they only allowed 47 stolen bases… less than half of the league average. So in some ways, one could argue that TEX catchers (primarily Rodriguez) prevented 57 stolen base attempts. Now, of course, a CS is more valuable than a runner just being held on, since it adds an out and pulls someone off a bag. But when measuring the individual skill of the catcher, I think you have to take into consideration the number of times opponents just said, “nevermind.”
The more I think about it, the more I think there should be a WOWY component. But that said, there needs to be an accounting for people being unwilling to even TEST Pudge’s arm. I would think that SB/LgAvgSBA would do the trick. Again, you can (and perhaps SHOULD) account for WOWY stuff.. but I am also certain that the person behind the plate often discourages even SB attempts, so runners only go in the very best of circumstances, which artificially inflates the SB%.
You’ve got a good point, and it shouldn’t be difficult to compare stolen base attempts per opportunities.
I realized another thing that could alter these numbers for battery’s with low attempts. The runners that DO attempt steals against these guys are probably better than average base stealers, so the sample of attempts may not be league average. It’s a little like “vs LHB” stats for left handed pitchers that dominate lefties. Usually the lefty batters they face are the very best hitters and the mediocre lefties are benched for platoon righties. This is just my hypothesis, but it makes sense.