from Amazon
Tuesday, April 12, 2011
Hamilton College Baseball wOBA, FIP
I took their stats, and added some new metrics.
1. First, I put an estimate for reaching base on errors. Those are bad things that happened, and we don’t care if the pitcher was responsible or not. We just care who was on the mound when bad things happened. I figured each pitcher’s reaching on error as 8.5% of all non-hit, non-K batting outs.
2. I needed an estimate for plate appearances, and that’s at bats, walks, hit batters, and sac flies.
3. I needed a wOBA formula. That’s 0.89 “pluses” for walks and hit batters, 0.99 for singles and reaching on error, 1.21 for doubles, 1.39 for triples, and 1.62 for HR.
4. I needed a FIP formula, and that was 3 points for a HR, 1 for a walk or hit batter, negative one for a strikeout. We divide all that by PA, multiply by 47, multiply by 0.6, and we add 9.29.
5. LOB% is simply one minus runs allowed divided by runners on base.
Here’s how they look:
LOB% OBP wOBA BABIP FIP PA Jersey Pitcher
60% 0.379 0.380 0.380 7.47 124 5 Kevin Prindle
57% 0.406 0.407 0.365 9.70 69 3 Alex Zeig
72% 0.430 0.437 0.442 7.97 107 14 John Wulf
43% 0.464 0.467 0.425 10.50 140 30 Dan Kroening
63% 0.479 0.478 0.446 10.45 73 20 Joe Wagner
54% 0.490 0.482 0.471 9.87 49 17 Michael Affuso
42% 0.516 0.531 0.490 11.49 64 16 Alex Potoczak
53% 0.548 0.534 0.480 12.93 31 29 Tom Moriarty
50% 0.543 0.542 0.500 12.42 81 9 Colin Henneberger
100% 0.250 0.223 0.000 16.34 4 24 Justin Atwood
33% 0.429 0.410 0.333 13.32 7 22 Brendan Rafalski
67% 0.500 0.462 0.250 18.69 6 15 Joe Buicko
8% 0.765 0.717 0.667 20.90 17 31 John Summa
53% 0.465 0.465 0.432 10.28 772 TOTALS
We see a few things:
1. Prindle, Zeig, and Wulf have so far had the best performances. But of those three, Wulf is the one that has stranded a ton of runners. So, if you are looking to understand why his team is 3-0 with him on the mound, while they are 4-13 otherwise, that plays a big part.
2. Prindle’s FIP is similar to Wulf, but Prindle’s BABIP is a very low .380 (compared to the team total of .432).
When you put the two together, we see that Prindle gives up fewer hits, but Wulf pitches better with men on base. But both pitchers have similar K, BB rates. The end result is Wulf ends up with fewer runs scored than Prindle. Chances are, we’re going to see both pitchers end up similar for the rest of the year, as Prindle gives up more hits, but pitches better with men on base.
What would be cool to see is the pitch-by-pitch data.
3. Also note the correlation between FIP and BABIP. The r is a very high 0.70 on only 61 BIP! So, to all those people who think that pitchers have little control on batted ball: that’s not true. See, the difference is that in MLB, all the pitchers are pretty good, and have been selectively sampled to be pitchers who don’t get hit hard. Basically, all the pitchers in MLB have a true talent of .280 to .320, and when talent is that tight, well, it’s hard to find the signal from the noise. If you see someone with a .290 BABIP on 500 BIP, well, that could have come from a .310 pitcher, or a .280 pitcher. It’s just too hard to tell.
But at Hamilton College, the spread in pitching talent is very wide. Instead of the true talent being .280 to .320 like in MLB, it’s more like .380 to .520. With that much spread in talent, you simply need alot fewer observations in order to see the talent signal from the sample size noise. Furthermore, with such a strong link to FIP, it reduces our need to rely on observed BABIP solely to find the true talent BABIP.
Good luck to Prindle, Wulf, and the rest of the pitchers at Hamilton. I’ll check back in when the year is over!
Recent comments
Older comments
Page 1 of 344 pages 1 2 3 > Last »Complete Archive – By Category
Complete Archive – By Date