The Elusive Clutch Hitter–Part II
I made some mistakes, some careless, some unknown, with the charts included in my post titled “The Elusive Clutch Hitter,” and I wanted to clear them up.
The first correction shows the batters in my sample who had an increase of at least 10% in batting average with runners in scoring position (RISP) vs. no runners in scoring position (nRISP). It was my intention to do this all along–indeed, I had posted earlier and woke up the next day and realized I had compared RISP batting average with career batting average, which would cause an overlap in the data. I trashed that post prior to it being published, but it appears I made the same mistake, at least with this particular chart. Here’s the corrected chart:
 |
| From FanGraph Graphs |
The columns on the left are nRISP situations, and the right are RISP situations. with the difference calculated as (rispBA-nrispBA)/(nrispBA). If I’ve amalgamated the data correctly (and I made every attempt, but errors will happen), adding the nRISP and RISP AB and hits columns should equal the player’s career totals. I have no idea how I got the years wrong in the initially published chart, but the same essential data is presented.
This next chart shows those players who had at least a 10% decrease in RISP batting average. This chart is essentially unchanged from the initial one, other than also making sure I was comparing RISP vs. nRISP hitting and not comparing to career hitting.
 |
| From FanGraph Graphs |
The only thing I’ll add here is something I forgot to state in my original post, which is that I’ll go so far as to say that most of these hitters are sample size issues, since most of these are current players.
The other thing I failed to mention in my original post is that all errors are mine, although as I amalgamate the mounds of data available from the various sorts, I try my hardest to be sure I’m accurate. Thanks for all the comments so far.
Great post, but I do have a problem with the definition of “clutch”. If a batter knocks in a runnner in scoring position without ending the inning, he produced in the clutch. I know that in certain situations with a runner on third, as a hitter, I would intentionally try not to do too much with the ball and would intentionally fly out to deep in the outfield to score a run. So I would sacrifice a hit to score a run or win the game. That was clutch hitting. I think you may want to analyze different data to see if someone is clutch or not.
if he can intentionally hit a deep flyball to score a run, how abt intentionally hitting a single, or better a homerun? dude when he hit a sac fly, he actually failed
Kampfer,
If you manage to get a run in, how is it you failed? The entire point of baseball is to score more runs than another team. You advanced a runner into score, you succeeded. Especially, when its the game winning run. Also, it is a lot easier to hit a long fly ball than to hit a single, double, triple, or homerun. You only have to hit a ball to a deep part of the park with little or no care for whether a fielder will catch a ball. To get a hit, the best hitters of all time failed 6 out of ten times they put a ball into play (i.e. a 400 batting average).
I have to side with Thurston on this one. If you can get the run, you’ve produced in the clutch. For that matter, you can also get a clutch walk. I once saw Bobby Abreu back on the Phillies bluff his way into walking the winning run in. He sat there with a 3-2 count, looked intense, gripped the bat like he was going to hit a grand slam… then just laid off, took the walk, won the game.
This is probably why people often use WPA to determine clutch hitting. With that said, there’s still definitely room to look at clutch performance. I would tend to think that a batter increases WPA, regardless of the magnitude, he’s performed in the clutch. That might mean hitting a grand slam or it might mean just knocking a runner over for the next guy to bat in.
Correction. Not just any increase in WPA, clearly it needs to be a RISP situation. Sorry if I didn’t make that clear. I felt it was assumed, but realized I didn’t mention it.
Looks like most of the commenting has ended, so I want to add some concluding thoughts:
1. Thanks for the interest, kind words and suggestions. This is the beginning, not the end of this line of research for me.
2. I am fully aware of the shortcomings of using batting average. I am currently in the process of combing through the 2010 box scores, and have made my way through about 30+ games, so I have a pretty good sample. These numbers are rough, but indicative:
3,180 RBIs
528 were done making an out (6 bunt outs, 3 double plays, 9 fielder’s choice, 172 ground outs, 253 sac flys and 85 walks, none of which were intentional, thank God).
I just haven’t quite figured out what standard to use when viewing OBP and clutch hitting. I’ll play around with the data and see what comes up.
3. I’m also attempting to measure clutch pitching using similar metrics. I’m not at all sure what I’ll do about OBP in that regard, and am thinking instead of using a measure as follows:
Traditional OBP=(H+BB+HBP)/(AB+BB+HBP+SF)
Modified (H+BB-IBB+HBP+SF)/AB
I’ve run those numbers, I just don’t know if they mean anything.
4. In case I never made it clear before, I firmly acknowledge there are clutch situations. All I was attempting to study was whether a player’s ability to hit in the clutch varied from his ability to hit in general.
Thanks again, thanks to the nice folks at FanGraphs for putting up my corrected charts, and I hope to contribute again.
Scott, thanks to you for the article. It is pretty amazing the amount of work you are doing for a post. This is definately thought provoking.
It is very interesting that there are some players who put up significantly better results over a large amount of PA 3000+. Luck it seems would have nothing to do with these cases.
Not to make even more work for you, but the accepted methodology (which I learned about at Baseball Prospectus) is to split a career into even and odd numbered years and compare halves, if a hitter trends badly in one half, but very well in the other, he’s not really clutch, he’s just lucky sometimes, and unlucky others. If ‘clutch hitting’ were an actual skill, we would not be talking about it, because the guys who were blessed with it would stick out like a sore thumb. Players with high walk and homerun rates are displaying a skill, generally, and are able to call upon this skill year in and year out. Clutch hitters, not so much.