So this is my first fanpost here at McC, so any and all feedback is greatly encouraged and appreciated. Also, this is my first research project using PITCHf/x, so feedback is welcomed for that as well. All PITCHf/x data is taken from www.texasleaguers.com and all heat maps are courtesy of www.fangraphs.com. Now let's get this bad boy started, shall we?
As you probably saw, yesterday saw not one, not two, but three articles about Matt Cain and his xFIP-defying superpowers. It started with Paapfly's post "Matt Cain ignores xFIP, again and again", then got picked up by FanGraphs' Dave Cameron, Tom Tango, and SBNation's own (!!) Rob Neyer.
First, some statistics. For his career Matt Cain is the owner of 3.45 ERA, which is fantastic. His FIP, which strips luck and fielding out of the equation, gives him a slight bump to 3.84, which is still quite good. xFIP adjusts the home run per fly ball rate to the league average of 10.6%, which hurts Cain. A lot. His 7.0% career mark is a full 33% below the league average. The theory behind xFIP is that the HR/FB ratio is a statistic that a pitcher cannot control, and as such it fluctuates year to year, much like BABIP. The thing with Cain is, over the course of his career (almost 1,100 IP), that rate hasn't fluctuated. In fact, it has stayed remarkably constant. Here are his HR/FB ratios since 2006, his first full season as a starter: 7.1%, 5.5%, 6.8%, 8.4%, 7.4%. Still, Cain has largely been dismissed as a fluke or merely lucky, just the other day Baseball Prospectus writer Bill Baer called Cain "one of the most overrated pitchers in baseball." Baer continued to cite Cain's home ballpark, and its well-known penchant for suppressing homers. But, as Paapfly points out, Cain's career home/road splits simply don't show it. His home HR/FB ratio is 6.7%, on the road, its 7.4%. That's a result, but not much of one.
Let's try to tackle this problem using PITCHf/x. Some basics: Cain throws five pitches: a four-seam fastball that sits 91-92 and touches 95, a curveball that sits around 77 and has about 7 inches of downward break, a changeup that sits around 85 and breaks inward to a RHB, a slider that sits 85, and a seldom-used two seam fastball that sits 92-93. He throws his fastball most often, about 58% of all pitches. His most effective pitch, according to FanGraphs linear weights, is either his four-seamer or his changeup, both coming in at just over a run above average per 100 pitches. He decreased the usage of his slider this year, favoring the changeup.
By looking at Cain's PITCHf/x splits, when can glean some assumptions out of the data. The first obvious thing one notices is that Cain likes to stay away. This makes sense: Cain's fastball has at best average velocity, and off-speed pitches are generally more effective on the outside part of the plate because a hitter has more trouble turning on one.
Another thing that struck me is Cain's pitch splits by batter handedness. He throws his changeup 23.2% of the time to lefties, and only 8.7% of the time to righties. On the other hand, he throws his slider 16.1% of the time to righties, while only 2.1% of the time to lefties, even less than he throws his two seamer. His curveball and two-seamer usage are in the same ballpark, regardless of batter handedness.
All of this makes intuitive sense. Cain's changeup breaks about 7 inches toward his pitching arm side, or away from a left handed batter. Meanwhile, his slider is relatively flat, but looking at his heat map, he throws it almost exclusively low and away from a righty. This illustrates the way he uses what is probably his worst pitch to his advantage: it remains very hard to hit a low-and-away slider over the fence. The changeup heat map shows the same thing: Cain throws it exclusively low and away to lefties.
Here's Cain's slider heat map vs. RHB:
And here's his changeup heat map against LHB:
Now, on to the comparisons. Dave Cameron made some comparisons to pitchers with similarly low HR/FB rates, and found that even for pitchers that kept their HR/FB low for four consecutive seasons, as Cain has, their HR/FB rose about 2% in the next two seasons. As commenter DrBGiantsFan pointed out, Cameron's sample was primarily pitchers that were just reaching and flying past their peaks. Paapfly responded to Cameron's post, and part of that post pointed out that Cain also has a very high infield fly ball percentage, at 12.9% in his career. Cameron and Paapfly use Roy Oswalt as the best Cain comparison, but I think I've found one better: Jered Weaver of the California Angels. Look at their batted ball percentages: flyballs: 48.5% Weaver, 45.3% Cain; infield fly balls: 13.1% Weaver, 12.9% Cain, home run per fly ball: 7.9% Weaver, 7.0% Cain. Both pitchers have high flyball and infield fly percentages, but don't give up very many homeruns.
Briefly, Weaver is a right-handed pitcher with a five-pitch arsenal. He throws the same five pitches as Cain, but he uses his two-seamer and his slider far more often. An analysis of the his handedness splits shows what you were probably expected: he throws his slider 31.2% of the time against righties, but only 5.1% of the time against lefties. Conversely, he throws his changeup on 8.1% of the time against righties, but 18.9% of the time against lefties.
The similarity to Cain's splits is striking. Unlike Cain, Weaver also mixes up his curveball according to batter handedness: 25.3% of the time against lefties, but only 0.2% of the time against righties. That's right, Weaver threw exactly 3 curveballs to righties this year. Meanwhile, looking at Weaver's heat maps confirm their similarities.
Weaver's slider vs. RHB:
And his changeup vs. LHB:
So what does it all mean? Weaver and Cain approach each at bat similarly, and each get the same results: locate the fastball away, mix it up with sliders away to righties and changeups away to lefties, with the occasional curveball and two seamer. Each gets a lot of pop-ups, a lot of fly balls, and not a lot of home runs. I think this can be attributed to their superior command, as both walk very few batters: Weaver, 2.17 per 9, Cain, 2.43 per 9 in 2010.
Pitches that are breaking away from the batter are very hard to hit for home runs, but are very easy to hit lazily into the opposite field. Weaver and Cain are both adept at locating their pitches exactly where its hardest to hit strongly, and therefore induce a lot of weak pop flies to the opposite field. After the PITCHf/x analysis, I am convinced this is a repeatable skill for both Cain and Weaver, provided they continue their strike-throwing ways.