Clutchiness Breakdown

When I posted my article on Kosuke Fukudome yesterday, loyal reader VegasWatch pointed out that the Cubs outfielder’s opening day home run likely contributed the bulk of his 0.52 clutch score. Therefore, after being given the label of “clutch” the net sum of all of Kosuke’s clutchiness would not add up to much.

The formula for clutch, as defined in the glossary here, is:

Clutch = WPA/pLI – WPA/LI

For further clarification, pLI refers to the average leverage index of all game events for a given player while WPA/LI refers to context neutral wins; in other words, what the player produced regardless of the situation he entered into. This formula calculates the performance level of a player in crucial situations relative to his standard production. If a player has a .330 batting average in high leverage situations but hits .330 everywhere else, he is not considered clutch. This is not to say he lacks talent, but rather he just produces at a high level in all situations and isn’t necessarily stepping his game up in crucial plate appearances.

The Kosuke example made me wonder which other players were greatly benefiting from a big play. Looking at the top eight clutch scores before the stats updated last night, I tracked the biggest individual play for each of the eight and compared the clutch score of that singular play to the net sum of their other plays. This way we can see which player’s clutch labels are truly derived from one big play as opposed to those who have been a bit more consistent in stepping up. Here are the eight, with their overall clutch score and the three required components of their biggest play – note that the pLI refers to the season average, not the game average:

Pat Burrell (1.33): 0.899 WPA, 3.56 LI, 1.09 pLI
Melvin Mora (1.30): 0.418 WPA, 5.14 LI, 1.04 pLI
Freddy Sanchez (1.27): 0.363 WPA, 4.65 LI, 1.03 pLI
Skip Schumaker (0.93): 0.287 WPA, 4.29 LI, 1.04 pLI
Jeremy Hermida (0.86): 0.294 WPA, 2.61 LI, 0.94 pLI
Bobby Abreu (0.84): 0.512 WPA, 5.44 LI, 0.92 pLI
Manny Ramirez (0.81): 0.482 WPA, 2.38 LI, 0.95 pLI
Joe Mauer (0.80): 0.364 WPA, 4.35 LI, 1.07 pLI

With these figures, here is the breakdown of the big play clutch vs. the clutch in all other plate appearances:

Pat Burrell: 0.57 big play, 0.76 other
Melvin Mora: 0.32 big play, 0.98 other
Freddy Sanchez: 0.27 big play, 1.00 other
Skip Schumaker: 0.21 big play, 0.72 other
Jeremy Hermida: 0.20 big play, 0.66 other
Bobby Abreu: 0.46 big play, 0.38 other
Manny Ramirez: 0.30 big play, 0.51 other
Joe Mauer: 0.26 big play, 0.54 other

Pat Burrell had the most clutch “big play” when he hit a walkoff two-run home run against the Giants on May 2nd. However, according to these numbers, Abreu actually benefited the most from his play; he is the only one whose big play exceeded the net sum of all other clutch plays.

On the flipside, Freddy Sanchez and Melvin Mora have been very consistent in raising their performance level in high leverage situations. When talking about a player’s clutchiness, though, it really only takes one or two big plays to cement the label. We could remove the one big play and look at all other performances but since one play can change a fan’s perception of clutchiness that just would not be fair; regardless of whether or not the clutch benefits from a huge play or a group of smaller plays added together, the bottom line is that these players have helped their team win games by stepping up in crucial situations.

Print This Post

Eric is an accountant and statistical analyst from Philadelphia. He also covers the Phillies at Phillies Nation and can be found here on Twitter.

6 Responses to “Clutchiness Breakdown”

You can follow any responses to this entry through the RSS 2.0 feed.
  1. Eric, could you explain how you got the big play clutch scores? I’m not getting the same numbers as you are. I’m seeing Howard’s big hit worth about .55 “clutch”, and for Mora I’m getting about .32 “clutch”.

    I’m just calculating his clutch score minus the one big play, or I think doing (WPA / avg(pLI for the season)) – (WPA/LI) to find the value of an individual play would work too.

    Vote -1 Vote +1

  2. Eric Seidman says:

    David, I see where the error is and I’m correcting it now. What I had done was WPA/(pLI – WPA/LI) instead of (WPA/pLI) – (WPA/LI). Everything should make more sense now.

    With pLI I am looking at the individual game in which the big play took place. So, Burrell had a 0.899 WPA on the big play, with a 3.56 LI, and in that May 2 game, he had a 1.38 pLI.

    When plugged in, (WPA/pLI) – (WPA/LI) = 0.6514-0.2525 = 0.40.

    So, that play accounted for a 0.40 clutch, and the net sum of all other plays, entering last night, would be 0.93, since he had a 1.33.

    Vote -1 Vote +1

  3. Eric Seidman says:

    With Mora:

    .418 WPA/1.92 pLI in the game of his big play = .2177
    .418 WPA/5.14 LI of the big play = .0813

    WPA/pLI – WPA/LI = ..136, or .14

    Vote -1 Vote +1

  4. I would use the pLI for the season when looking at an individual clutch play instead of for just that game since clutch isn’t calculated on a per game basis and then added up. Ideally you’d want to pretend like that play never happened and then calculate what his clutch score would have been without it entirely.

    You’re probably going to be undervaluing the “clutch” aspect of the play a little to considerably by using just that games pLI because chances are it’s going to be a much higher pLI than his “normal” pLI would be.

    For instance, if a player had only one plate appearance in that game and his LI in that appearance was let’s say 5, and then he hit a home run worth .6 wins, you’d be calculating his clutch score for that plate appearance at 0.

    Vote -1 Vote +1

  5. Eric Seidman says:

    Okay, gotcha’, we’re all clear on that. I’ll go in and re-calculate.

    Vote -1 Vote +1

  6. Eric Seidman says:

    All numbers corrected, my apologies to anyone for the confusion.

    Vote -1 Vote +1