Clutch Pitchers (January 28, 2004)
Here are your clutch and non-clutch pitchers:
Pitcher Clutch PA LI
Jones, Todd -45 1218 1.73 (best)
Benitez, Armand -42 1199 1.87
Mesa, Jose -37 1306 1.66
Weathers, David -37 1390 1.19
Alfonseca, Anto -35 1219 1.75
Colon, Bartolo -32 3562 0.97
Mlicki, David -32 2583 0.96
Johnson, Randy -30 4102 1.01
Looper, Braden -29 1307 1.18
DeJean, Mike -28 1191 1.06
Lieber, Jon -26 3446 0.97
Fetters, Mike -26 815 1.12
Baldwin, James -26 3063 0.93
Glavine, Tom -25 3842 0.99
Leskanic, Curt -25 998 1.20
Zimmerman, Jeff -25 927 1.43
Sauerbeck, Scot -24 1148 1.06
Trachsel, Steve -23 3226 0.96
Remlinger, Mike -23 1232 1.42
Mulholland, Ter -23 2063 0.89
...
Lowe, Derek 20 2058 1.31
Yoshii, Masato 20 2481 0.86
Trombley, Mike 20 1041 1.36
Guthrie, Mark 20 966 0.97
Rapp, Pat 21 2157 0.95
Halama, John 21 2413 0.97
Nitkowski, C.J. 21 1135 0.94
Adams, Terry 22 1937 1.18
Woodard, Steve 23 1958 0.87
Tapani, Kevin 23 2140 0.97
Rivera, Mariano 23 1066 1.96 !!!!!
Herges, Matt 24 1273 1.16
Biddle, Rocky 25 1008 0.92
Millwood, Kevin 25 3202 0.97
Milton, Eric 25 3355 0.96
Rupe, Ryan 26 2051 0.96
Padilla, Vicent 28 1304 1.16
McElroy, Chuck 30 878 0.69
Timlin, Mike 33 1224 1.28
Acevedo, Juan 38 1377 1.23
Wasdin, John 46 975 0.64 (Worst)
John Wasdin is an interesting case. Not only is he the worst clutch pitter in the league, but he also has the 2nd lowest LI in the league. That is, he pitched in one of the fewest high pressure situations in the league. To be able to amount that amount of poor clutch performance in so little opportunity is an incredible feat.
And as with the hitters, the spread of performance among the pitchers was exactly as would have been expected if a clutch talent did not exist. Again, that doesn't mean clutch talent does not exist, but rather, I can't find it using this very rigorous process.
--posted by TangoTiger at 11:48 PM EDT
Posted 11:55 p.m.,
January 28, 2004
(#1) -
tangotiger
Btw, I redid the LI for all my relievers, but this time, I used the win probability table for 1999-2002 (as opposed to my generic 1974-1990 win prob table).
Boy, did that make a difference. It's alot easier to win a game with a 3-run lead in 1986 than in 2000. As a result of redoing the LI, Troy Percival comes in with a 2.17 for 1999-2002. 9 pitchers had an LI above Bruce Sutter's career LI of 1.90. Keith Foulke came in #23, which shows how horribly he was used. Lou Pote brought up the rear with pitchers with at least 800 PA with .54, and the incredible John Wasdin had a .64 for 2nd to last place.
Posted 2:15 p.m.,
January 29, 2004
(#2) -
FJM
Wasdin's "achievement" isn't merely interesting; it appears (to me at least) to be mathematically impossible. Same goes for McElroy, who finished exactly one notch above Wasdin in both LI and the non-clutch standings. That in itself is suspicious. Moreover, both pitchers actually had winning records during the period in question (12-11 for Wasdin, 8-4 for McElroy). That's really hard to believe, if they both gave up a lot of important runs (despite being used in very few high leverage situations). Compare that to the next 2 pitchers on your list (Rupe and Acevedo), who allowed 58 clutch runs between them and posted a combined 35-62 record. But the kicker is this: Wasdin allowed a total of 124 earned runs. Throw in some inherited runners and unearned runs and he's probably somewhere around 150. So to be 72 runs worse than the average pitcher in the clutch, just about all the runs he allowed would have to have come in high LI situations, which we know is impossinble. Could you double-check your numbers for these 2 pitchers?
Posted 2:44 p.m.,
January 29, 2004
(#3) -
tangotiger
I wanted to run Wasdin, because I couldn't believe it either. Here we go.
This is how Wasdin did, in various pressure situations (the more negative, the better):
PA avgLI runs
582 0.18 (19)
198 0.73 (7)
84 1.24 10
69 1.90 7
42 3.35 8
So, he had 582 PAs in low-pressure situations, and he was great. He had 42+69 high pressure PAs, and he was terrible.
The total of his performance was 975 PAs, for a total of -1 runs. That is, without quantifying the pressure of his performance, he was a league average pitcher.
However, those 582 PAs had very little pressure (18% of normal swing impact). Recasting the above numbers by multiplying by the LI, and we get:
PA runs
103 (3)
145 (5)
104 13
131 13
140 28
His total leveraged PAs is now 622. (Remember, his total was 975, so his LI is .64).
The total of his runs is now +46 runs.
Hmmm... how did I get 72? Looks like I have a bug in my program, as I double-counted his LI (46/.64 = 72). I'll fix that shortly.
So, Wasdin is 47 runs worse, based on the game context, that his random performance would have dictated.
Looking at Wasdin's 42 PA with the highest pressure (i.e., with the game most on the line), this is what he did:
HR: 5
2B: 3
RBOE: 2
1B: 5
HBP: 2
BB: 4
Outs: 21
Now, that is complete futility. 21 times on base in his 42 most PAs where the game was most on the line. That's a .500 OBA. Each of these PAs had an LI of at least 2.5 (which is fireman territory).
John Wasdin pitched exactly as bad as Barry Bonds hits well, and he did it with the game on the line.
If I go back to his 111 PAs with the LI of at least 1.5, and the hitters got on base 49 times in 111, or .441, with a total of TWELVE HR. Remember, that's in 111 PAs.
A truly horrible performance.
Posted 2:53 p.m.,
January 29, 2004
(#4) -
tangotiger
FJM: thanks for spotting the bug. List has now been updated. Seeing Mariano Rivera as "bad clutch" and Armando as "good clutch", I will present their numbers in a few minutes as well.
Posted 3:02 p.m.,
January 29, 2004
(#5) -
tangotiger
Mariano Rivera. Here we go. 312 PA (I exclude all IBB) with an LI of at least 2.5
6 HR, 6 2B, 6 RBOE, 52 1B, 1 HBP, 17 BB, 224 outs.
That's an OBA of .282, which is excellent.
However, overall, he had 1066 PA, with an OBA of .260, which is sensational. So, his "poor" showing with the pressue for Mo is that he went from super-great overall to great overall.
I'll do Armando next.
Posted 3:06 p.m.,
January 29, 2004
(#6) -
tangotiger
Armando. .275 OBA overall. .220 OBA with the game on the line!
So, Armando went from being great to unfreakingbelievable.
Benitez had 353 PAs with the game on the line. This is what he did:
8 HR, 2 3B, 8 2B, 2 RBOE, 28 1B, 2 HBP, 28 BB, 275 outs.
Posted 3:51 p.m.,
January 29, 2004
(#7) -
tangotiger
When I look at my "Win Advancement" (WA), who do we see bringing up the rear? John Wasdin. He had 9.7 WA and 14.2 LA (loss advancements). That makes him -4.5 wins compared to average. This WA process was done on a PA-by-PA basis.
As you can see, my current "rough" process detailed in this thread had Wadin at -46 effective runs. At a 10:1 runs to win converter, that translates to -4.6 wins compared to average.
Posted 4:35 p.m.,
January 29, 2004
(#8) -
FJM
Your welcome. There are some surprising names at the top of the Clutch list, particularly Weathers and Mlicki. Could you break down their performance by LI category, as you did with Wasdin? I wonder if they are mirror images of him: great when the pressure is on, lousy when it isn't.
Posted 4:57 p.m.,
January 29, 2004
(#9) -
tangotiger
"effOBA" is LWTS converted into an OBA, as described in a thread a few weeks ago.
Mlicki
PA avgLI effOBA
600 0.31 0.407
1018 0.77 0.374
561 1.23 0.353
356 1.87 0.346
48 3.11 0.258
Weathers
PA avgLI effOBA
489 0.21 0.334
284 0.73 0.321
183 1.24 0.340
263 1.97 0.293
171 3.48 0.243
Should I just generate this for all the pitchers?
Posted 7:10 p.m.,
January 29, 2004
(#10) -
FJM
If it isn't too much trouble, I'd appreciate it. Mlicki's clutch performance looks like nothing more than a statistical fluke, resulting from just 48 very high LI PA's. (He does seem to be somewhat better under moderate to heavy pressure as opposed to no pressure at all, though.)
Weathers is much more interesting. He looks like a closer-in-waiting. He's actually pretty good all the time, but he really steps it up when the pressure is heavy to extreme. If you were to assume his overall performance is the mean, how much would you regress his high pressure stats?
Posted 11:16 p.m.,
January 29, 2004
(#11) -
tangotiger
Pitcher stats are alot less reliable than hitter stats. If you take his last 2 lines of pressure (434 PA), you would regress that by a little less than 50%. If he was a hitter, you'd regress by about 30%.
But, this regression is towards the population mean.
However, since we "know" that that observed distribution of high-pressure OBA compared to overall true talent OBA is exactly what you'd expect from luck, you should regress everyone's high-pressure OBA by 100% towards the true talent.
So, either you regress his performance by 50% towards the pop mean of .340, or you regress his performance by 100% towards his true mean.
Bottom line, our best guess is that all these performances are just random variations centered around the player's true mean.
Posted 2:26 a.m.,
January 30, 2004
(#12) -
FJM
We could give it more weight if we knew it was coming from improvements in SO, BB or HR as opposed to a BIP.
Posted 10:37 a.m.,
January 30, 2004
(#13) -
Mike Green
This interesting article covers one type of "clutch pitching"- victory-important pitching. Another type of clutch pitching- run-important pitching, is also interesting because it explains in some situations (over at least a season) disparities between DIPS ERA and actual ERA. Pitchers who in a season have substantially worse performance with runners on than with no one on will end up with a significantly worse actual ERA than DIPS ERA.
This was true of Cory Lidle's 2003 season, and I am guessing that if you studied the 10 pitchers with the highest differential between DIPS and actual ERA, at least three or four of them will have this attribute.
Incidentally, knowing what we know of John Wasdin from this study, if you were a Manager at the All-Star break, and your team had just acquired Wasdin after he had a fine half-season starting at Triple A in 2003, would you put him in to start straight off against the Yankees and Red Sox? Somehow, I suspect Leo Durocher would have had him in mop-up relief until the Devil Rays or better yet the Tigers came to town and given him some strong liquid medicine before the game to settle him down.
Posted 11:35 a.m.,
January 30, 2004
(#14) -
Brian P
Did anyone catch the number of Phillies SP's at the bottom of the clutch list? Millwood, Milton AND Padilla all rank in the bottom 8! It's doesn't look like the LI's are that great either.
Posted 11:45 a.m.,
January 30, 2004
(#15) -
tangotiger
Having a low LI jsut means that they weren't facing many "fire" situations. We don't know why that is. Is it because their teams took early leads? Is it because they pitched great, to let their teams score? Is it because they pitched horribly, and their teams couldn't keep up the score?
LI is not an indication of the pitcher, so much as the situation. LI = level of fire.
As for Wasdin, to the extent that clutch pitching might exist, Wasdin is by far the worst clutch pitcher around. If a psychologist were to interview Wasdin, and tell us that Wasdin is perfectly fine, and whatever intangible characteristic they can figure out, then that's another nail in the "clutch doesn't exist".
Wasdin is the poster boy here, as much as Patrick Roy is the poster boy for great clutch performer. I BELIEVE that Roy does have something extra.... up until the point in the semi-finals when he gave up 9 goals against Detroit. So, who knows.