Ju Think Ju Know Who is Teh Clutch??

[bigh0rt]

It's funny when reputation is true for once. This year the top three clutch hitters were Pujols, Ortiz and Jeter. Using play by play data and the Leverage Index (which gauges how critical an AB is to win probability) these are the top three. All of these AB's had a Leverage Index of 3.5+ which basically means the game was on the line at the time these AB's occured. Which means it was late in the ballgame and the teams best chance to win the game. Basically, if you take the last three innings of a ballgame and you have a 1-2-3 8th and 9th inning while in the 7th inning having bases loaded and 2-out, then the 7th inning is the 'clutch' inning because it was your last 'best chance' to win the game.


Pujols has been really good in the clutch:

According to Fangraphs, Pujols leads the majors in 'clutchiness'. Last year, when he won the MVP, he was actually -2.16 in clutchiness WPA, showing just how fluky clutch hitting can be.

Perhaps the best thing to come out of WPA is something called Leverage Index (LI), which was invented by Tangotiger and measures the criticality of each potential play. An average LI is 1.00, so an LI of three is three times more critical to the outcome of a game than average.

With LI, you can look at plate appearances an entirely different way. For instance, here are Pujols's top 15 plate appearances ranked by LI (all plate appearances with an LI over 3.5):

Date Opp Inning Outcome LI WPA
2006-08-1 CHN 9 Flied Out 6.90 -0.179
2006-04-1 CIN 9 Home Run 5.40 0.646
2006-04-1 CIN 8 Single 4.71 0.075
2006-09-1 ARI 9 Flied Out 4.40 -0.108
2006-05-0 COL 8 Home Run 4.39 0.380
2006-06-0 CHN 7 IBB 4.32 0.024
2006-04-2 PIT 9 Single 4.27 0.283
2006-04-0 CHN 5 Single 3.97 0.222
2006-07-1 LAN 10 IBB 3.88 0.005
2006-08-2 CHN 8 IBB 3.83 0.025
2006-09-0 PIT 9 Fielder's Choice 3.66 -0.104
2006-07-0 HOU 12 Flied Out 3.63 -0.117
2006-04-2 CHN 5 Single 3.61 0.212
2006-07-0 KC 11 Single 3.60 0.135
2006-05-2 HOU 7 Home Run 3.57 0.501

In his 15 most important plate appearances, Pujols has gone 8 for 12 with three home runs and three intentional walks, for a total WPA of 2.0, almost exactly his clutchiness score. Clutch hitting may come and go, but when it comes it can win pennant races.

Of course, so has David Ortiz:

I kind of like that table, so let's do another. Boston's 'Mr. Clutch', David Ortiz, has had 14 plate appearances with an LI of 3.5 or higher. Here are the outcomes of each one:

Date Opp Inning Outcome LI WPA
2006-05-2 NYA 8 Struck out 6.05 -0.155
2006-09-0 KC 8 Single 5.76 0.496
2006-04-2 TB 9 Walk 5.27 0.195
2006-07-3 CLE 9 Home Run 4.83 0.785
2006-07-0 CHA 13 Struck Out 4.61 -0.078
2006-06-2 PHI 12 Single 4.44 0.383
2006-07-2 LAA 11 Single 4.44 0.383
2006-05-2 TB 5 Double 4.18 0.338
2006-05-0 BAL 6 Double 4.00 0.296
2006-08-2 LAA 9 IBB 3.96 0.030
2006-06-1 TEX 9 Home Run 3.69 0.904
2006-08-0 TB 10 IBB 3.69 0.013
2006-05-0 TOR 9 Struck out 3.60 -0.095
2006-08-0 CLE 9 Flied out 3.60 -0.095

Not too shabby, eh? Ortiz has been 7 for 11 in these plate appearances, with two home runs, a walk, and two intentional walks for a total WPA of 3.4. Despite his protestations, however, playing time and the Red Sox's collapse will keep the MVP trophy out of Ortiz's hands this year.

Speaking of the Red Sox's collapse, check out the August WPA totals for Boston. Check out Mike Timlin!

Derek Jeter hasn't been shabby in the clutch, either:

The shortstop MVP candidate in New York (no, not Jose Reyes, unfortunately) hasn't been too bad in high-LI situations either.

Date Opp Inning Outcome LI WPA
2006-05-1 TEX 9 Double 5.40 0.427
2006-05-2 NYN 9 Ground out 5.02 -0.146
2006-04-1 KC 8 Home run 4.71 0.624
2006-07-2 TOR 7 Walk 4.70 0.173
2006-08-2 BOS 9 Single 4.60 0.289
2006-08-1 BOS 7 Double 4.54 0.427
2006-06-2 ATL 9 IBB 3.88 0.005
2006-05-2 NYN 8 Single 3.72 0.139
2006-04-2 TB 10 Single 3.69 0.084
2006-06-0 BAL 9 Single 3.69 0.344
2006-06-1 OAK 7 Struck out 3.61 -0.108
2006-06-0 OAK 9 Ground out 3.60 -0.095
2006-06-2 ATL 12 Ground out 3.60 -0.095

In 13 plate appearances with an LI over 3.5, Derek Jeter has been 7 for 11 with a home run and an intentional walk. His WPA in those situations has been 2.2. Jeter has a slight lead over Joe Mauer and Santana in WSAB and his MVP candidacy may hinge on what you think of his fielding prowess (or lack thereof). SG, for instance, looks at Jeter's MVP qualifications in more detail at the Replacement Level Yankee Blog.

So, how many posts until we hear a certain New York third baseman mentioned? The Over/Under is set at 3.

[mak1277]

David Wright? I bet he was in the top 15-20%.

[Another Blown Save]

David Wright!
Oh wait...

[Half Massed]

Interesting, I wonder which "unknown" type players have turned out to be surprisingly clutch. A Marco Scutaro type for example.

[ayebatter]

Interesting, I wonder which "unknown" type players have turned out to be surprisingly clutch. A Marco Scutaro type for example.

I'm bettin' Freddy Sanchez with 200 hits, but only 6 hrs shows up fairly high

[acsguitar]

Pretty cool stat I guess.

Although Clutchiness is great but what about the players who get their team on the board first yet fail at the end.

[The Miner Part 2]

Pretty cool stat I guess.

Although Clutchiness is great but what about the players who get their team on the board first yet fail at the end.

cough jim thome cough

[Tavish]

The biggest problem I have with WPA (or more precisely LI) is that it appears to vastly overweight late inning events over everything that comes before. A pitcher can get a Win while going 2 runs over 7 innings and receive basically the same WPA as both the guy who comes in to pitch a scoreless 8th and the closer who pitches a scoreless 9th. I'm all for giving different weights to events (hell I'm one of the biggest L-Weights boosters around here) but LI has always gone a little overboard IMO.

http://www.hardballtimes.com/main/article/ten-things-i-didnt-know-last-week31/

[Coppermine]

Is Andy Van Slyke on the list?

[RynMan]

Nice post h0rt. I think the best way to measure clutch is as they have done. Use win expectancy data with play by play data.