Why Josh Tomlin Gives Up Homers by Eno Sarris October 28, 2016 Right-hander Josh Tomlin starts for Cleveland tonight in Game Three of the World Series. While he does a lot well, he also has a weakness — namely, that he gives up more homers than the average pitcher. It’s possible that, among the explanations you’ll hear regarding that weakness, most will relate either to how it’s because Tomlin lives in the zone or never gives in or something along those lines. He certainly doesn’t walk people, so there would seem to be some logic to that argument. It’s also tempting to point to the relationship between his walk rate and his home-run rate because of the extremes he’s reached in both departments. Record-setting extremes, actually. This year, Tomlin gave up 16 more homers than he did walks. In over 5200 qualified starting-pitcher seasons since World War II, nobody has ever produced a greater discrepancy in that department. Only four times — Carlos Silva in 2005, Brian Anderson in 1998, Brad Radke in 2005, and then Tomlin this year — has that difference run into double digits. Still. The walk rate is nice. And it’s probably not why he gives up homers. Run walk rate against home-run rate and you find that there’s really just no relationship, statistically. (The p-value is .11 and the r-squared is .004, if you’re wondering.) Even if there were, there’d be quite a bias in that relationship — if you give up home runs and walk a lot of guys, you’d have to be amazing at striking guys out or you’d be in trouble and out of baseball fairly quickly. Different pitchers can live in the zone, though — especially with the fastball — and ultimately record starkly different walk rates depending on the other offerings in their repertoire. Tomlin was in the zone only 52.6% of the time this year, which was actually a below-average figure among his peers, so it might be wrong to say that he “lives in the zone” at all. But he’s produced higher zone figures in the past: in 2011, for example, he recorded a 59.3% zone rate on his fastball, a figure that sits among the top 6% since 2010. So, do pitchers that live in the zone with the fastball necessarily give up homers? No. It’s semi-compelling to see that Phil Hughes in 2014 had the second-highest fastball zone percentage of our sample, and he seems to be a Josh Tomlin type, but there’s Addison Reed, too, leading the entire sample with his effort this past season and not exhibiting any sort of homer problem at all. Cliff Lee was always in the zone with his fastball. Clayton Kershaw this past season recorded the 14th-highest fastball zone percentage since 2010, among 1366 qualified pitcher seasons, and things seem to be working out alright for him. There’s no relationship between fastball zone percentage and home runs per batter faced; the p-value (.356) and r-squared values (.0014) are terrible. This may seem hard to square with something we know is true about baseball: power comes more often from inside the zone than out. Look at the isolated power zone for the league, and this is true. Duh: batters like it middle-middle. Two things may be going on here. One: we may not be asking the question with sophisticated enough tools. Or, two: if a pitcher can get the fastball in the zone, it’s also likely that he can hit those blue and white spots in the zone and avoid the red. Command and control aren’t the same thing — some pitchers can throw into the zone generally but throw it middle-middle too often — but if a pitcher’s throwing the fastball in the zone a ton, perhaps it’s actually a sign both of command and control. In any case, this past season Tomlin hasn’t actually been around the zone that often. Not judged by overall zone rate (44.9% for Tomlin, 44.6% for the league) or by fastball zone rate (52.7% for Tomlin, 53.2% for the league since 2010). So that’s not it, probably. Could it be count-based? Perhaps Tomlin is more predictable in certain counts? There’s some evidence that he’s got a problem on 2-1 counts over his career. Tomlin vs MLB: Home Runs per Pitch by Count HR/P by count Tomlin League All 1.13% 0.72% 2-0 1.02% 1.02% 2-1 1.82% 1.04% Tomlin: 293 2-0 counts, 494 2-1 counts. It’s maybe anecdotally interesting Tomlin has given up homers — relative to league average — much more often in 2-0 counts than 2-1 counts. The sample isn’t huge, but it’s not tiny either. Maybe he’s a little predictable in 2-1 counts, in a slightly unconventional way: Josh Tomlin Career Pitch Usage by Count Pitch All Counts 2-0 2-1 Fourseam 38% 54% 36% Sinker 6% 10% 6% Change 9% 6% 10% Curve 14% 1% 4% Cutter 33% 29% 43% SOURCE: Brooks Baseball He throws the cutter in 2-1 counts more than in any other count. Relative to the league, though, the cutter is his best pitch by whiffs. And though it gives up more line drives per ball in play than his other pitches, across all counts it’s given up fewer home runs per pitch than any offering other than his curve. In 2-1 counts, the home-run rate on that pitch doubles. All of his home runs on 2-1 counts this year were off the cutter. Maybe he’s a little predictable there, but this still only explains four of his 36 homers on the season. Tomlin did have a mediocre swinging-strike rate this year, 6.1% to the league’s 9.5%. Here, we might have a bit of a relationship. The p-value is okay (.0044), meaning that these two variables are probably not related out of chance; however, among the 579 qualified pitching seasons since 2010, swinging-strike rate explained only 1.4% of the variance in home runs per batter faced. It’s not a strong relationship. Whiffs can help you suppress home runs, but just by a tiny bit. There is one way that Tomlin was a standout this year. Lowest Fastball Velocities, 2016 Name Fastball Velo R.A. Dickey 82.3 Jered Weaver 84.0 Doug Fister 87.0 Josh Tomlin 87.8 Marco Estrada 88.1 Hisashi Iwakuma 88.2 Dallas Keuchel 88.3 Dan Straily 89.1 Zach Davies 89.3 CC Sabathia 89.4 Qualified Starting Pitchers Only And his fastball velocity does have an effect on home-run rate. With a p-value under .0001, and an r-squared of .075, it’s not the strongest effect, but it’s the strongest effect we’ve seen so far, and if you pair it with swinging-strike rate, you see an even stronger relationship. Velocity and stuff do help suppress home runs, to an extent. This isn’t an unprecedented finding. The relationship between fastball velocity and homers has been shown before. It’s probably also why Tomlin has more than doubled his curveball usage in the postseason at times. It’s an approach that’s both less predictable and less fastball-focused. Maybe it’s rude to point out that Tomlin’s stuff is probably why he gives up home runs, and so you might hear that it’s about how often he’s in the zone, or his walk rate, or the like. But it really looks like, despite an okay cutter and a decent curve, it’s that slow fastball that is probably the biggest reason for all of the homers.