Getting and Not Getting the Calls: Final 2012 Results by Jeff Sullivan October 9, 2012 All the way back in May, I came up with a pretty simple way to calculate “expected strikes” based on data available at FanGraphs. I don’t know if I was the first person to do this, and it’s so simple I’d be surprised if I were, but I remember me so I’m linking to me. Once you have expected strikes, you can compare that total to actual strikes, and maybe then you can learn something about the pitcher(s) or the catcher(s) or about something else. I”ll explain further! FanGraphs provides for you total pitches, total strikes, and plate-discipline data based on PITCHf/x data. By using zone rate, you can come up with pitches in the zone, which leads to knowing pitches out of the zone, which leads to knowing swings at pitches out of the zone. Based on those numbers, you can end up with an expected strikes total. You’re way ahead of me — I probably don’t need to explain this in great detail. The reason I’m re-visiting this now is because the 2012 regular season is over, meaning those numbers aren’t going to change anymore. Meaning now it’s time for a 2012 regular season in review, according to this nameless statistic. I’m interested in the difference between actual strikes and expected strikes per 1,000 pitches, and that’s what you’re going to find in the tables below. This is a post that requires very little in the way of text since the tables tell you just about everything you need to know. The first table shows all of the teams in baseball. A positive Diff/1000 value means the team got more strikes than PITCHf/x thought it deserved. A negative Diff/1000 value, therefore, means the team got fewer strikes than PITCHf/x thought it deserved. After that table, you’ll find more tables, with individual pitchers. These are the top and bottom 10 pitchers in Diff/1000, given a minimum of 50 innings. Got it? You got it. Note that the overall league average is about -5, and not zero. That is, for every 1000 pitches, there were five fewer strikes than PITCHf/x thought there should’ve been. Off we go! Table 1: Team Data Team Diff/1000 Brewers 11 Braves 11 Reds 5 Yankees 5 Cardinals 2 Rays 2 Diamondbacks 2 Giants 1 Nationals 0 Phillies 0 Red Sox -1 Mets -2 Padres -2 Astros -3 Angels -4 Cubs -5 Orioles -7 Blue Jays -8 Athletics -10 Marlins -10 Tigers -10 Royals -10 Rockies -10 White Sox -11 Dodgers -12 Rangers -13 Twins -13 Indians -17 Mariners -18 Pirates -19 Table 2: Top 10, Individual Pitchers Name Diff/1000 Brad Ziegler 40 Livan Hernandez 39 Chad Durbin 27 Sean Burnett 26 Matt Albers 25 Randall Delgado 25 Craig Kimbrel 24 Scott Atchison 23 Yovani Gallardo 22 Freddy Garcia 21 Table 3: Bottom 10, Individual Pitchers Name Diff/1000 Jeff Gray -39 Charlie Morton -36 Justin Masterson -34 Tony Watson -33 Hector Santiago -32 Joel Hanrahan -31 Jeanmar Gomez -30 Samuel Deduno -30 Adam Ottavino -30 Jose Valverde -29 You might be curious, so here’s the difference between Brad Ziegler and Jeff Gray, visualized, courtesy of Texas Leaguers. First Ziegler, then Gray. So! Is it perfect? Of course not. I don’t know what PITCHf/x considers to be the strike zone, and there are differences between that strike zone and real-life strike zones. What this is is basically free of bias, since PITCHf/x is objective. So while you should treat the numbers with care, it seems to me there might well be something in there. Per 1000 pitches this season, Brewers pitchers got 11 more strikes than expected, and Pirates pitchers got 19 fewer strikes than expected. Brewers and Pirates pitchers threw many thousands of pitches. I bring this back to pitch-framing because it seems like a good and reasonable explanation. It doesn’t explain everything, but I think it probably explains a lot of everything. Mike Fast’s research on the subject was rather fond of Jonathan Lucroy, Brian McCann, and David Ross, among others. It wasn’t so big on Rod Barajas or Carlos Santana. The Mariners’ placement makes a lot of sense considering all of their innings were caught by Miguel Olivo, John Jaso, and Jesus Montero, and those guys are all supposed to be bat-first, except in the case of Olivo, who is more like nothing-first. The results probably aren’t shocking, especially because I’ve posted partial-season results before that looked pretty similar. There are a lot of things at play, here. Some pitches are just more difficult for umpires to identify correctly than others. Surely, within this data there is noise. But within this data there is probably also signal, and I find that to be most interesting. Getting borderline strike calls is not the most important thing, but it is a thing, and sometimes a critical thing. And when it comes to getting borderline strike calls, not all teams and pitchers are the same.