Benefit of the Doubt: Mo and the Wide Zone

by J-Doug on Feb 9, 2011 10:00 AM EST in Pitch f/x Goodies

Last week, we looked at the pitchers who cost or saved their teams the most runs via their respective strike zones. This week, I present to you the pitchers whose zones made the biggest difference over approximately 9 innings.

When we calculate the LWD differential for every 75 pitches called (the average number thrown in a 9 inning matchup), we find Wide Zone Mo at the top of the list. In regular season games from 2008-2010, Mariano Rivera's zone (heat map above) has saved the Yankees about 1.61 runs every 9 innings. This works out to about 0.16 runs per outing.

Considering how seldom Rivera allows any runs at all, the runs saved by his strike zone are especially valuable.

On the other end of the scale we find maligned reliever Andrew Miller. His zone (mapped below) cost the Marlins about 0.52 runs per 9.

Below is a list of the ten most pitcher- and hitter-friendly zones over the last three seasons based on LWD/75. Same as last week, only pitchers with at least 250 pitches called in each of the previous three seasons qualify.

Pitcher	Called Pitches	LWD	LWD/Year	LWD/75	Standard Deviations
Rivera, Mariano	1404	-30.19	-10.06	-1.61	-1.73
Hoffman, Trevor	997	-13.82	-4.61	-1.04	-1.07
Hernandez, Livan	4877	-67.49	-22.50	-1.04	-1.07
Wilson, Brian	2029	-25.41	-8.47	-0.94	-0.96
Lowe, Derek	4965	-54.63	-18.21	-0.83	-0.83
Moyer, Jamie	4055	-38.96	-12.99	-0.72	-0.70
Cordero, Francisco	1878	-17.79	-5.93	-0.71	-0.69
Okajima, Hideki	1499	-13.90	-4.63	-0.70	-0.68
Halladay, Roy	4985	-43.49	-14.50	-0.65	-0.63
Peavy, Jake	3051	-25.41	-8.47	-0.62	-0.59
~	~	~	~	~	~
Madson, Ryan	1445	5.43	1.81	0.28	0.45
Lindstrom, Matt	1210	4.69	1.56	0.29	0.46
Hernandez, Felix	5054	20.06	6.69	0.30	0.46
Perez, Oliver	2971	12.44	4.15	0.31	0.48
Linebrink, Scott	1352	5.75	1.92	0.32	0.49
Nippert, Dustin	1813	8.38	2.79	0.35	0.52
Rodney, Fernando	1758	8.39	2.80	0.36	0.53
Robertson, Nate	2718	14.35	4.78	0.40	0.58
Ohlendorf, Ross	2618	14.41	4.80	0.41	0.60
Miller, Andrew	2110	14.55	4.85	0.52	0.72

Felix Hernandez, who brought up the rear in last week's column, has only the 8th worst zone among qualifying pitchers.

In this week's list, I also present standard deviations from the mean (since, when talking rate stats, this actually means something). As you can see above, only three of the best pitcher's zones—and none of the best hitter's zones—fall more than a standard deviation away from the mean. This would imply that the real strike zones for most pitchers (excepting Livan Hernandez, Trevor Hoffman, and Mariano Rivera) are not significantly different from the average zone from 2008-2010.

This is somewhat misleading. As it turns out, the rankings of hitter's and pitcher's zones are somewhat sensitive to the minimum called pitch cutoff. Let's take a look at another list for comparison's sake. In this chart, a pitcher qualifies if he threw 750 called pitches over the last three seasons, regardless of whether he threw a minimum of 250 in any single year.

Pitcher	Called Pitches	LWD	LWD/Year	LWD/75	Standard Deviations
Rivera, Mariano	1404	-30.19	-10.06	-1.61	-1.73
Duchscherer, Justin	1407	-22.44	-7.48	-1.20	-1.25
Hoffman, Trevor	997	-13.82	-4.61	-1.04	-1.07
Hernandez, Livan	4877	-67.49	-22.50	-1.04	-1.07
Mussina, Mike	1814	-23.03	-7.68	-0.95	-0.97
Wilson, Brian	2029	-25.41	-8.47	-0.94	-0.96
Weathers, David	1200	-14.55	-4.85	-0.91	-0.92
Gallardo, Yovani	3445	-41.71	-13.90	-0.91	-0.92
Martin, J.D.	938	-11.08	-3.69	-0.89	-0.89
Lowe, Derek	4965	-54.63	-18.21	-0.83	-0.83
~	~	~	~	~	~
Cabrera, Daniel	2173	11.29	3.76	0.39	0.57
Robertson, Nate	2718	14.35	4.78	0.40	0.58
Rupe, Josh	987	5.35	1.78	0.41	0.59
Baez, Danys	923	5.04	1.68	0.41	0.59
Ohlendorf, Ross	2618	14.41	4.80	0.41	0.60
League, Brandon	1182	6.83	2.28	0.43	0.62
Lowe, Mark	1298	7.95	2.65	0.46	0.65
Samardzija, Jeff	811	5.08	1.69	0.47	0.66
Miller, Andrew	2110	14.55	4.85	0.52	0.72
McGowan, Dustin	903	22.36	7.45	1.86	2.25

Under the new criteria, we find a number of new pitchers populating the list. Perhaps most notable is Dustin McGowan, whose LWD/75 of 1.86 is beyond one standard deviation from the mean.

In fact, there's a notable if not earth-shattering year-to-year correlation for the stat: LWD/75 over the 2008-2009 period predicted about 45% of the variance in 2010 LWD/75. So we can say that the stat is a somewhat reliable indicator of what the real zone for any individual pitcher is over the course of a season. This is true even if very few individual zones ever deviate by more than one sigma from the mean.

PitchF/X data originate from Darrell Zimmerman's SQL-based PitchFX database, run expectancy data by Tom Tango at Inside the Book. Photos courtesy of Wikimedia Commons and licensed under Creative Commons terms.

Previous episodes in the Benefit of the Doubt series:

Tweet Comment 2 comments | 0 recs |

Do you like this story?

Comments

I am not convinced

that the differences among pitchers is actually due to a more or less favorable zone (as if we knew the exactly 3D location and movement across the plate of every pitch with no uncertainty) rather than the movement of a pitcher’s repertoire of pitches as well as the percentage of their pitches that are near the boundaries of the zone.

For example, if a pitcher threw pitches that were only in the middle of the zone or well outside the zone, you would expect that his overall zone would be nearly equal to the average MLB zone. For a pitcher who threw all of his pitches on the edges you would expect his zone to be larger or smaller, depending on whether he had good or bad control, even if the umpire was almost perfect with his calls.

Another example is this: If a pitcher has a lot of movement on his pitches, you might expect him to get fewer correct calls on strikes, again, through no conscious bias by the umps.

My examples may be wrong, but I think my concept is clear and is this: I don’t necessarily think that we can tell from this data which pitchers actually get the benefit of the call or not, unless we look at each type of pitch individually and compare it to the league average for that same type of pitch. And even then, there might be differences depending on sequence, location, etc.

by mgl on Feb 10, 2011 12:28 AM EST reply actions