Breaking Numbers, Pt. 2

Only one more off-day left before baseball resumes in the Yankee-Red Sox Time-Space Continuum. Our second and last look at All-Star break rivalry-centric leaderboards off the beaten trail continues with a look at the pitchers. I chose more stats mostly because there are more interesting stats for pitchers than there are for hitters. At least, that's my opinion.

We'll start with a couple stats purporting to provide a better overall look at a pitcher than straight ERA. Mostly we'll stick to the eight principal starters, but in some cases relievers also have qualified or been added.:

ERA+

  1. Josh Beckett, 140
  2. Jon Lester, 121
  3. A.J. Burnett, 117
  4. CC Sabathia, 114
  5. Tim Wakefield, 108
  6. Joba Chamberlain, 104
  7. Brad Penny, 99
  8. Andy Pettitte, 91

Wins Above Replacement

  1. Josh Beckett, 3.5
  2. Jon Lester, 3.4
  3. CC Sabathia, 2.9
  4. Tim Wakefield, 2.1
  5. Brad Penny, 1.8
  6. A.J. Burnett, 1.4
  7. Andy Pettitte, 1.0
  8. Joba Chamberlain, 0.9

ERA+ simply takes ERA and adjusts it for league average and ballpark, as opposed to WAR, which looks more at peripherals, and we see the difference: Burnett has not pitched as well as his ERA suggests, and neither has Chamberlain. Penny has pitched better.

Pitching Runs Created
  1. Josh Beckett, 62
  2. Jon Lester, 58 
  3. CC Sabathia, 58 
  4. A.J. Burnett, 51
  5. Tim Wakefield, 45 
  6. Jonathan Papelbon, 36 
  7. Andy Pettitte, 36 
  8. Brad Penny, 33 
  9. Joba Chamberlain, 32 
  10. Ramon Ramirez, 28 
This is a Hardball Times stat. No Mo? He's 11th, with 26. How is this possible? Well, it's based on ERA with adjustments for league average, luck and defense, with a further adjustment to put it on the same baseline as Batting Runs. Papelbon still has a sterling ERA, but there's no way he's been as good as Rivera.
 
Again, ERA-based stats reward lucky pitchers. Here's another set of peripheral-based numbers:
Fielding Independent Pitching
  1. Jon Lester, 3.32 
  2. Josh Beckett, 3.36 
  3. CC Sabathia, 3.73 
  4. Tim Wakefield, 4.26 
  5. Brad Penny, 4.28 
  6. A.J. Burnett, 4.65 
  7. Joba Chamberlain, 4.84 
  8. Andy Pettitte, 4.95 

FIP-ERA

  1. Jon Lester, -0.55 
  2. Brad Penny, -0.43 
  3. CC Sabathia, -0.13 
  4. Tim Wakefield, -0.04 
  5. Josh Beckett, 0.01 
  6. Andy Pettitte, 0.10 
  7. Joba Chamberlain, 0.60 
  8. A.J. Burnett, 0.88 

Like the difference between ERA and WAR, we see that Burnett and Chamberlain join Pettitte at the bottom of the list, and that all three have been the beneficiaries of good luck or good defense. On the other hand, Lester and Penny in particular are well outperforming their ERAs, and a lot of this has to do with the Sox' lousy left-side defense, particularly in Lester's case. To wit:

Defense Efficiency Ratio (Pct of non-HRs turned into outs behind pitcher)
  1. CC Sabathia, .735
  2. Josh Beckett, .722   
  3. A.J. Burnett, .717 
  4. Tim Wakefield, .708 
  5. Andy Pettitte, .694 
  6. Brad Penny, .681 
  7. Joba Chamberlain, .674 
  8. Jon Lester, .667 

DER is basically reverse BABIP. If league-average BABIP for a pitcher is around .290-.300, then league-average DER behind him should be .700-.710.

So we see an anomaly in that Chamberlain's FIP is actually much higher than his ERA, but that his DER is lower than average. Lester is exactly where we would expect. He has been disproportionately affected by bad luck and bad defense so far this season.

Another factor in affecting ERA is giving up hits at bad moments, some combination of failing in the clutch and bad luck. WPA addresses this by looking at the pitchers who have contributed the most to their team's victories based on each individual play.

Win Probability Added
  1. Josh Beckett, 1.76
  2. Tim Wakefield, 0.86
  3. Jon Lester, 0.55
  4. A.J. Burnett, 0.30
  5. CC Sabathia, 0.17
  6. Brad Penny, -0.48
  7. Andy Pettitte, -0.81
  8. Joba Chamberlain, -1.27

Pitching well in close games will get you a lot more WPA than pitching well on either side of a blowout. So dividing out the Leverage Index of each play gets you a more context-neutral figure:

WPA/Leverage Index (context-neutral WPA)

  1. Josh Beckett, 2.07
  2. CC Sabathia, 1.76
  3. Jon Lester, 0.54
  4. Tim Wakefield, 0.42
  5. A.J. Burnett, 0.07
  6. Joba Chamberlain, -0.76
  7. Brad Penny, -0.78
  8. Andy Pettitte, -1.06

Chamberlain, Penny and Pettitte are as a group one pitcher from the bottom among AL-qualified starters, suggesting none of them is as good as we may think he is. This list is probably how we'd intuitively rate the pitchers this season, with the exception of Burnett and Wakefield. Burnett by all these measures has been performing much worse than his numbers imply.

Pitches per Plate Appearance
  1. Tim Wakefield, 3.5 
  2. CC Sabathia, 3.8 
  3. Josh Beckett, 3.9 
  4. Andy Pettitte, 3.9 
  5. Joba Chamberlain, 3.9 
  6. A.J. Burnett, 3.9 
  7. Brad Penny, 4.0 
  8. Jon Lester, 4.1 

Fastball Velocity

  1. Josh Beckett, 94.3
  2. A.J. Burnett, 94.2
  3. CC Sabathia, 94.1
  4. Brad Penny, 93.7
  5. Jon Lester, 93.5
  6. Joba Chamberlain, 92.4
  7. Andy Pettitte, 88.7
  8. Tim Wakefield, 72.7

Of the 10 fastest throwers in the league, five are on the Red Sox or Yankees (a caveat applies in that pitchers throwing cutters misclassified as fastballs are poorly represented on a list like this).

Ultimately those numbers are totals of how pitchers have performed in individual games. I like the following two stats because it gives a better idea of how dominant a pitcher can be in any given performance:

Quality Start Pct.

  1. Josh Beckett, 72
  2. Tim Wakefield, 65
  3. A.J. Burnett, 65
  4. Jon Lester, 61
  5. Brad Penny, 53
  6. CC Sabathia, 47
  7. Joba Chamberlain, 47
  8. Andy Pettitte, 44

Average Game Score

  1. Josh Beckett, 57
  2. CC Sabathia, 55
  3. Jon Lester, 55
  4. A.J. Burnett, 54
  5. Tim Wakefield, 50
  6. Joba Chamberlain, 48
  7. Brad Penny, 46
  8. Andy Pettitte, 46

QS% is based on the traditional definition of a quality start (6 IP/3ER), which is the only way I can explain how Sabathia has managed a quality start in just 47 percent of his appearances, while averaging a game score of 55 (where 50 is a quality start). Likely, Sabathia has several starts where he pitched seven or eight innings and gave up four earned runs, earning a game score above 50 but being disqualified from a traditional quality start. Wakefield, as usual, is racking up a large number of minimum-level qualty starts, as is Penny. These confirm that Beckett by any definition is the dominant ace of these two ballclubs.

Of course, individual games hinge on how well a pitcher executes each individual pitch.

FanGraphs uses Pitch f/x to break down the expected run values of every single count and how a pitcher's pitches changed those counts and expectations, For example, if a pitcher has a 1-0 count, then throws his fastball for a strike to make the count 1-1, his fastball is credited with the difference in runs now expected from the new count. Similarly, the difference in run values from a 3-1 count with one out and the bases loaded turned into a ground ball double play is credited to the changeup thrown to induce the grounder. As a result, you get a look at the overall effectiveness of a pitcher's different pitches.

wFB/C (Linear-weights fastball value in runs per 100 pitches)

  1. Tim Wakefield, 1.57
  2. CC Sabathia, 0.65
  3. Josh Beckett, 0.21
  4. Brad Penny, 0.14
  5. Jon Lester, -0.46
  6. A.J. Burnett, -0.80
  7. Andy Pettitte, -1.06
  8. Joba Chamberlain, -1.20

Wakefield of course almost always throws his fastball in counts where a hitter is likely to be taking all the way (3-0 and first pitch), so he piles up the strikes with it. Given Wakefield has the slowest fastball in the league by 10 mph, maybe hitters should stop taking those 3-0 pitches. Surprisingly, Lester and Chamberlain see poor results. Chamberlain's almost certainly is a result of poor command. Lester's may still be recovering from his poor start, or this may reflect that his fastball is straight enough to really get hammered when he leaves it over the plate.

wCB/C

  1. A.J. Burnett, 1.88
  2. Josh Beckett, 1.66
  3. Jon Lester, 1.57
  4. Andy Pettitte, 1.24
  5. Joba Chamberlain, -0.07
  6. Tim Wakefield, -0.49
  7. CC Sabathia, -3.66
  8. Brad Penny, -3.79

No surprise that Burnett, Beckett, Lester and Pettitte all feature filthy curveballs. Pettitte's having the worst season of the eight starters, and it would be worse if not for this one pitch. Sabatahia and Penny need to put the curveball away for a while.

wCH/C

  1. CC Sabathia, 3.48
  2. Josh Beckett, 2.55
  3. A.J. Burnett, 0.69
  4. Jon Lester, -0.24
  5. Andy Pettitte, -0.90
  6. Joba Chamberlain, -3.18

Wakefield and Penny are the only two pitchers in the league to never throw a changeup (or a pitch classified as a changeup, at any rate). Lester's managed to make his new pitch value neutral in his first season of really using it. Not bad. Beckett's present near the top for all three of these pitches indicates what we all know — he can and does dominate with three plus pitches, something no other pitcher in either rotation can say.

So what about the results themselves?

Strike Percentage

  1. Justin Masterson, 65
  2. Josh Beckett, 65
  3. C.C. Sabathia, 64
  4. Brad Penny, 64
  5. Jon Lester, 64
  6. Tim Wakefield, 63
  7. Andy Pettitte, 61
  8. A.J. Burnett, 60
  9. Joba Chamberlain, 59

First-Pitch Strike Percentage

  1. Josh Beckett, 64
  2. Brad Penny, 61
  3. A.J. Burnett, 60
  4. Justin Masterson, 60
  5. CC Sabathia, 59
  6. Jon Lester, 58
  7. Andy Pettitte, 57
  8. Joba Chamberlain, 56
  9. Tim Wakefield, 53

For some reason, Justin Masterson is considered eligible for these lists (he's apparently faced 3.1 hitters per team game). We all can see that Chamberlain's struggles have been with commanding the strike zone, and we see that here: Below 60 percent of his pitches for strikes, and the first pitch is a strike only 56 percent of the time. Sabathia's first-pitch strike percentage is also surprisingly low, while Penny scores surprisingly well in both categories.

Not all strikes are created equal of course. Home runs are considered strikes. Here's some ways of seeing how well a pitcher is fooling hitters:

Outside the Zone Swing Pct.

  1. Jon Lester, 27.8
  2. CC Sabathia, 27.6
  3. Andy Pettitte, 26.5
  4. Josh Beckett, 25.6
  5. Joba Chamberlain, 23.5
  6. Tim Wakefield, 21.7
  7. A.J. Burnett, 21.3
  8. Brad Penny, 19.1

Lester and Sabathia are very good at inducing swings on balls. Burnett and Penny not so much.

Zone Contact Pct.

  1. CC Sabathia, 83.8
  2. Jon Lester, 85.8
  3. Tim Wakefield, 87.4
  4. Josh Beckett, 87.4
  5. Brad Penny, 90.2
  6. Joba Chamberlain, 90.9
  7. A.J. Burnett, 90.9
  8. Andy Pettitte, 91.8

Similarly, Lester and Sabathia are very good at inducing swings and misses on pitches in the strike zone. Chamberlain, Burnett and Pettitte all struggle with this, and that's why the rest of their numbers are equally poor (relatively speaking).

Finally, here are a pair of stats I just included for fun:

Run Scored Pct. With Runner on 3rd, Less Than 2 Outs

  1. Jonathan Papelbon, 17
  2. Ramon Ramirez, 25
  3. Edwar Ramirez, 29
  4. Daniel Bard, 30
  5. Josh Beckett, 38
  6. Manny Delcarmen, 38
  7. Tim Wakefield, 38
  8. Alfredo Aceves, 43
  9. Joba Chamberlain, 50
  10. Mariano Rivera, 50

This is more curiosity than anything. For relievers, the sample sizes are tiny (Rivera has allowed one of his two runners to score, Papelbon one of his six. You could argue you'd actually rather have the guy who only gets into two of those situations rather than six of them). But by luck or clutch or what have you, Beckett and Wakefield have been uniquely effective as a starter in preventing runners from scoring in a situation that otherwise is arguably as optimal as any in baseball for run production.

Stolen Base Pct. Against

  1. Brad Penny, 94
  2. Josh Beckett, 90
  3. Tim Wakefield, 88
  4. Joba Chamberlain, 76
  5. Andy Pettitte, 67
  6. A.J. Burnett, 64
  7. Jon Lester, 60
  8. CC Sabathia, 54

The grouping of this ranking (Sox in the top three, with the lefty the only pitcher with a respectable number) suggests an indictment of the catchers, as opposed to the pitcher. That Wakefield isn't first on this list seems an indication that George Kottaras may be better than he gets credit for.

That's all, folks. Hope you enjoyed it!

12 comments… add one
  • very cool

    dw (sf) July 16, 2009, 1:28 pm
  • I’m really happy that you posted these on a non-baseball day. I don’t have the time right now, but I’ll read it later! Bring back baseball!

    Atheose July 16, 2009, 1:40 pm
  • Does the ERA+ ballpark adjustment for this year or? I would’ve expected a bigger boost just because the stadium’s been HR-happy.. (which might not be reflected in things like FIP because it doesn’t affect BABIP because they are HR’s..)

    Lar July 16, 2009, 1:49 pm
  • wow. amazing work paul.
    for the record, re cc., he qs percentage is lower than it might be because a couple of early starts he was allowed to remain in a couple of batters too long, and gave up late runs. so really, those go on girardi. as cliff corcoran notes over on bb today, you take out those anomalies, and the one short start with the phantom bicep injury, and his numbers are a lot better. which is to say, look out in the 2nd half.

    YF July 16, 2009, 2:16 pm
  • @Lar – I don’t think “they” (whomever the elusive they are) have updated ball park factors for the new Yankee universe home run stadium yet. :)
    so if you’re in a front office – with the pelthora of stats available, what do you base a decision on – who to go after, who to offer what to?

    dw (sf) July 16, 2009, 2:26 pm
  • you take out those anomalies, and the one short start with the phantom bicep injury
    …and the one where he didn’t get enough sleep the night before, and the one where the bullpen allowed two inhereted runners to score, and the one where he just didn’t have it, and the one where he had a fight with his wife right before the game…
    Taking out an injury-related start is valid, I think, but removing starts because of a manager’s decision starts us down a road in which we’re parsing the exact circumstances surrpunding every non-excellent start for every pitcher to explain why every start in the second half will be excellent. I’m pretty sure Sabathia will be the victim of bad luck or poor managerial choices in at least a couple starts in the second half, like all pitchers are.

    Paul SF July 16, 2009, 2:29 pm
  • “That Wakefield isn’t first on this list seems an indication that George Kottaras may be better than he gets credit for”
    Maybe I hallucinated this but I always though Wake was surprisingly not easy to steal against because of his short step to home.

    Nick-YF July 16, 2009, 3:25 pm
  • Yeah, NYS makes some of these stats as they apply to Yankee pitchers (an hitters, albeit to a significantly lesser extent) tough. Park effects when applied to these stats are usually based on three-year averages (for sample size and because the effects can show quite a bit of variance year to year).

    Mark-YF July 16, 2009, 5:30 pm
  • Pitcher, home, road OPS allowed
    Sabathia, .677, .628
    Burnett, .775, .696
    Chamberlain, .847, .697
    Pettitte, .865, .787
    Beckett, .536, .730
    Lester, .685, .717
    Wakefield, .756, .698
    Penny, .763, .849
    I’m not sure how ERA+ and OPS+ factor in a park adjustment for a park that has doesn’t even have a year of data yet. It’s certainly possible that the ERA+ (and other stats using park factors) may need to be adjusted up slightly (and OPS+ down slightly) to compensate for YS going from a 100 park factor to something like a 106-108.
    I’m not sure it would make much of a difference, and as we’ve established, NYS is really no more an overall hitter’s haven than Fenway is, and it’s obvious from the stats above that pitchers can and do succeed in that park.
    What’s especially odd, though, is that the Sox’ starters, except for Wakefield, perform significantly better at Fenway than on the road. One-year fluke? Intentional product of roster construction? Something they’ve figured out about how to pitch there? Worth thinking about.

    Paul SF July 16, 2009, 5:57 pm
  • I don’t know about you, but those stats suggest the Sox pitchers have been coached about how to pitch in Fenway. The one guy who has done worse at home is the one guy that doesn’t throw to spots.
    It will be interesting to see if the Yankee pitchers learn to pitch in the new park. The difference of course though is that Fenway has center field and right-center for pitchers. Neither are safe in the Bronx Bandbox.

    Rob July 16, 2009, 6:06 pm
  • The one guy who has done worse at home is the one guy that doesn’t throw to spots.
    Exactly what I was thinking.
    I’m excited to see that Lester has been especially unlucky. Most of the times I’ve watched him he’s been fantastic, and I’m excited for the 2nd half.

    Atheose July 16, 2009, 8:53 pm
  • Just had to say nice work and thanks for the good read!

    Chris July 16, 2009, 10:40 pm

Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.