April 9, 2014
Max Scherzer and the Sabermetric Approach to Pitching to the Count
Tuesday night, Tigers against the Dodgers, third inning, this sequence:
In the stats reports that the Angels generate for their minor-league pitchers, there’s a column that tracks 1-1 strikes. What you can measure you can manage, the McKinsey truism claims, and the 1-1 column helps the Angels encourage and emphasize count control.
First-pitch strikes are commonly counted, of course. Announcers mention them constantly. If you click “more stats” enough on Baseball-Reference, you’ll find a column that lists first-pitch strike rates for each pitcher. "The best pitch in baseball is strike one," is something you’ll hear a dozen times this year.
But “Of all twelve (counts),” Eric Seidman wrote in 2008, “Greg Maddux considers the 1-1 count to be of utmost importance. Missing on a 1-1 count shifts the momentum back toward the hitter whereas a successful 1-1 pitch can move the count’s favor further in the direction of the pitcher. The 1-1 count brings with it a run expectancy of -0.012 from the batter’s perspective; a ball shifts it to +0.037 whereas a strike causes a jump to -0.079. Maddux is right.”
So imagine a pitcher throws 65 percent strikes. Imagine all pitchers throw 65 percent strikes. Imagine that most pitchers throw 65 percent strikes in all situations, but that there is some random fluctuation here and there. So all pitchers throw 65 percent strikes but some pitchers throw 68 percent strikes on 1-1 and 64 percent every other time; and some throw 62 percent strikes on 1-1 and 66 percent every other time. If we believe pitches thrown on 1-1 are of more importance than the rest, then the 1-1 strike-throwers would get better results than the 1-1 ball-throwers. If it’s a repeatable skill, then it’s a valuable skill. If it’s not a repeatable skill—if the difference between the two groups of pitchers is truly random fluctuation—then one group got “lucky.” Just as sequencing of baserunners allowed can create “luck” for some pitchers, the sequencing of good and bad pitches might.
This article is not one of those “We looked at every pitcher in baseball history” inquiries that Russell Carleton is always talking about. This looks at one pitcher: Max Scherzer.
Last year, 66 percent of Scherzer’s pitches were strikes or put in play. League average was 64 percent, so he was a moderate strike thrower. But on 1-1 pitches, 74 percent of his pitchers were strikes or put in play. That’s the highest rate in baseball for a starter (followed by Charlie Morton, Michael Wacha, Derek Holland, Homer Bailey). The eight percentage point delta between his baseline and his 1-1 rate is the third highest in baseball (behind Morton and Jacob Turner, ahead of Holland, Hyun-Jin Ryu, and Stephen Strasburg). For whatever reason, Scherzer was a particularly good pitcher when, according to Maddux and pitching coaches everywhere, it mattered most. How did he do this?
First, there is some indication that, with Scherzer, this success on 1-1 is a skill. Over the past five years, he has had higher strike rates on 1-1 than overall each year—sometimes by small margins, sometimes large: