Clutch Hitting Study
By David Grabiner
I posted a previous article on my clutch study; I have now re-done it,
with everything properly weighted. The basic conclusion didn't change;
I still don't have evidence that clutch hitters exist, and if they do,
they cannot be very important.
I have 245 player-seasons to work with: every player who had at least
250 AB in the late innings of close games before the year in question
(either 1991 or 1992), and was listed as a regular in the Great American
Baseball Stat Book in that year.
The correlation between past and current clutch performance is .01, with
a standard deviation of .07. In other words, there isn't a significant
ability in clutch hitting; if there were, the same players would be good
clutch hitters every year.
The GABSB's definition of the late innings of close games is the seventh
inning or later with a one-run lead, a tie score, or the tying run on
base, at bat, or on deck.
The average player in my study lost 26 points of OBP in the clutch in
1991 and 1992, with a standard deviation of 116. 20 points was the
average loss for the 1984-1990 and 1984-1991 data, with a standard
deviation of 50.
I'm perfectly willing to accept that the correlation of .01 is real,
even if it isn't close to statistically significant. The linear
regression coefficient is .02, with a standard error of .15. That means
that Darren Daulton, the biggest choker in the major leagues over the
1984-1992 period (he lost 200 points of OPS), is expected to lose 24
points next year in the clutch instead of the normal 20. But that's too
small a difference even to measure; if he stretches a single into a
double in 100 AB, that would add 10 points to his clutch OPS, and 8
points to the difference.
If the correlation coefficient is actually .01, that means there is an
ability to hit in the clutch, with a standard deviation of about 8
points of OPS. I'm certainly willing to believe this, because there are
clear reasons why it should happen.
For example, Darryl Strawberry might be -16, one of the 2% of players
who are two standard deviations below the mean. He is a left-handed
hitter with a platoon split of 150 points of OPS; thus, if he faces 10%
more left-handed pitchers in the clutch than normal, that will cost him
15 points. For this to happen in 100 clutch at-bats, the opposing
manager would have to bring in a spot lefty to face him ten times.
Likewise, Mariano Duncan might be +16. He is a right-handed hitter
(formerly a switch-hitter who couldn't hit lefty) with a 200-point
platoon split in OPS. If he is due up to the plate 80 times in the late
innings of close games, which might normally be 48 times against
righties and 32 times against lefties, then he would gain 20 points if
he leaves for a pinch-hitter twelve times against righties and
pinch-hits four times against lefties.
And this is only one factor which affects clutch ability but has nothing
to do with bearing down in the clutch. Boggs, McGriff, and Strawberry
have 5 of the 14 worst choking numbers. All three are excellent
left-handed hitters with large platoon differentials; this makes them
vulnerable to spot lefties, because they do not leave for pinch-hitters.
Other possible factors include an ability to hit one-pitch fastballers,
who are often relief aces; an ability to hit the same pitcher better
when you see him for the third or fourth time in a game; and difficulty
seeing the ball in twilight.
