Randy Choate, Platoon Splits, and Arm Slots
It was the inning that shouldn’t have been.
First, in the tenth inning of Game Three of the National League Championship series, the Giants saw Brandon Crawford stroll to the plate against Randy Choate. It’s easy to say that the matchup didn’t favor the hitter based on Choate’s career splits. Choate has struck out 27% of the lefties he’s seen, and only walked 7.7%. Crawford walks 8.7% of the time against lefties, but his strikeout rate jumps to 24.5% when he’s seeing a southpaw.
Since both the expected strikeout rates were above average, you’d expect a multiplier effect. In other words: a lefty at the plate that’s particularly bad against lefties is even worse against a lefty that is particular hard on lefties. If you want to know more about this phenomenon, check out Steve Staude’s excellent post. If you want a table summary, check out this expected strikeout rate for a plate appearance based on the pitcher and batter’s respective strikeout rates:
| Pitcher’s K% | |||||
| 10% | 15% | 20% | 25% | ||
| Batter’s K% | 10% | 5.3% | 7.8% | 12.0% | 14.6% |
| 15% | 8.1% | 11.4% | 16.0% | 20.8% | |
| 20% | 11.8% | 15.6% | 20.6% | 26.9% | |
| 25% | 13.6% | 21.0% | 24.6% | 34.9% | |
Maybe we could have expected Brandon Crawford to strike out in about 35% of those at-bats. Or 31.3% if you use the calculator in the piece linked above. Add in the fact that Choate has allowed a .254 batting average on balls in play against lefties, and you realize that Crawford’s expected on-base percentage for that at-bat was probably around .200.
He wasn’t expected to do much against an arm slot like that — he admitted that it’s “hard to get comfortable with that arm slot — he’s almost coming from behind you.” Jeremy Greenhouse once represented platoon splits by arm slot, and you can see that where Choate releases the ball is particularly tough on lefties. (I’ve added the annotation.)
So yeah, the deck was stacked against Crawford. But at least he’d seen him before. As Crawford put it, “I’ve faced him a decent amount of times… not a whole lot of success, but I’ve actually been able to draw a few walks.” In ten plate appearances, Crawford has drawn four walks. Given the tiny sample, it’s not enough to move the statistical needle, but at least Crawford had some familiarity.
When the advantage swung around the other way, in the next at bat, it didn’t even seem like we’d ever get to see a true outcome between Juan Perez and Randy Choate. The bunt sign was on, as the home team was playing for one run and one run alone.
But it was the complete unfamiliarity that Perez had with the arm slot that ended up helping us to a non-bunt outcome. As Perez said after the game, he’d “never seen someone with a release point like that, where the ball is almost coming underhanded.” He’d certainly never bunted off of someone like that.
“After I failed at my job, I just wanted to make sure I hit the ball hard somewhere,” said. Perez. Now the math was in his favor. Now, given their respective strikeout rates against the hands at the plate, the expected strikeout rate for that plate appearance was 13.3%. Now, Perez had seen the ball twice from that arm slot, and now the batter was in the green.
After a foul and another ball, Choate had thrown 13 pitches. Not one had been a slider. Sinkers have the worst traditional platoon splits in baseball. Choate tried to throw an inside sinker and this is where it ended up.
Perez was ready for it. He had considered going the other way on the fifth pitch, on the outside, but when it was a ball, he recalculated. “I reminded myself he had a good sinker, so I choked up and looked to pull the ball,” he said after the game.
Maybe those ten plate appearances against Randy Choate prepared Brandon Crawford enough to beat the numbers stacked against him. Maybe those two failed bunt attempts prepared Juan Perez enough to pull the ball into left field and set the table for the end of the game. Those things we can’t know for certain.
Should the Giants have subbed a righty like Joaquin Arias in for Brandon Crawford? Should they have let Juan Perez hit away from the start? Should the Cardinals have taken Randy Choate out against a righty once a batter was on? Those answers seem a little more certain.
What we definitely know, though, is that Choate’s arm slot made a big difference in Tuesday night’s outcome. Even if there were things that shouldn’t have been.
With a phone full of pictures of pitchers' fingers, strange beers, and his two toddler sons, Eno Sarris can be found at the ballpark or a brewery most days. Read him here, writing about the A's or Giants at The Athletic, or about beer at October. Follow him on Twitter @enosarris if you can handle the sandwiches and inanity.
It seems a bit oversimplified to extrapolate from Crawford’s increased strikeout tendencies against LHPs that he is “particularly bad against lefties.” A lot of Crawford’s other numbers against southpaws don’t bear this out.
This is excellent and sound advice, and…
A pitcher’s own TTO splits have no predictive power. Cameron was using a strawman when he shows you Vargas’ own TTO numbers. What if they had been the reverse – he had gotten better each time through the order in his career? That would presumably negate his whole thesis, right? No. Because those personal, unique numbers are meaningless. We must make the assumption that ALL pitchers have a similar TTO penalty regardless of what their own numbers have been in one season or their career (pitchers with more or fewer pitches in the repertoire can be assumed to have a slightly less or more penalty).
What if a manager is going to bring in his closer – his ace – to face the opposing team’s heart of the order in a 1-run game in the 9th. But wait, he sees that with this pitcher he has done poorly against teams’ 3-4-5 hitters this season or in his career, even though he is a great pitcher overall. Would he not bring him in? No, because those numbers have no predictive value. By chance alone, SOME pitchers are going to have particularly bad numbers against 3-4-5 batters. SOME pitchers are going to have especially bad or good numbers against ANY subset of hitters or situations by chance alone.
What if you are thinking about throwing your best pitcher in an elimination game on Sunday, but wait, he has not so great numbers during the day, or on Sundays. Do you hold him out? No, same thing.
I’m going to keep harping on Cameron and other sabermetric writers until they stop implying and disseminating false and misleading information.
If we’re talking about the wording, I’ll concede the point. But if we’re talking about quoting a player’s strikeout rate versus lefties as being misleading… Strikeout rate stabilizes quickly and Crawford’s strikeout rate versus lefties has come in 497 plate appearances. It’s not like I quoted his OPS or wOBA against lefties is it?
And boom: your concise analytical evisceration of another reader’s pathetic prose has once again restored the cosmos to a state of snobbish equilibrium. Please, don’t get up, your next beer is on me, oh vaunted sabermetric hero.
So this comment was posted in at least three articles over a 3 hour span…is someone just trolling FG using MGL?
“Should the Cardinals have taken Randy Choate out against a righty once a batter was on?”
I’m amazed how much debate this question has prompted elsewhere (SBNation). Choate was the last lefty the Cardinals had available and had held LHH to an .093 BA this season, and after Perez (who can’t hit and was trying to sacrifice anyway) were 2 more LHH to come in Blanco and Panik. OF COURSE you stick w/Choate there. Why else would he even be on the roster if he’s not the best choice for precisely that situation?
Just an added tidbit for people who don’t know: Juan Perez is not a Grichuk type who actually has good platoon splits vs. lefties even though he’s not a good hitter. In fact he’s been an even worse hitter vs. lefties than righties (.501 OPS in 111 career PAs vs. LHP). Second guessing is always fun but I can’t see any reason other than hindsight why that particular decision should warrant any criticism at all.
I’ll weasel out of it because I wasn’t really focusing on it: I didn’t say *what* the answers were!
I know you weren’t engaging in that second guessing, Eno. I just saw enought of it elsewhere to think it was worth commenting on even though it wasn’t your focus.
BTW, where did you get the post-game comments from Crawford and Perez? Did you ask them the questions yourself? I don’t usually see beat writers and other mainstream media types asking about release points and such and it is really interesting to hear from the hitters how they factor that in to their approach.
Yeah I asked them those questions. It’s hard in the postseason, though. Not as much access.
Awesome stuff, Eno. Really. I can appreciate how tight access must be but please do your best to keep it up. The stuff you’re putting out right now marrying conventional reporting with statistical analysis and approaches so that we can read how the players themselves incorporate this information into their game is IMO the best baseball writing going on today.
Stop using a pitcher or hitter’s own performance v. RH and LH opponents to estimate their expected performance. It doesn’t work that way.
But, you are 100% correct. There was NO reason to take Choate out v. Perez when there was a 90% chance he was bunting, with 2 lefties following him. The suggestion that that might be correct is preposterous.
FYI, Choate has an enormous projected platoon split 63 points of wOBA. Juan Perez, 15 points, which is indeed quite low for a RHB (but NOT reverse – no RHB has a reverse expected platoon split!).
How are you getting this enormous projected platoon split for Choate if you’re not considering his platoon split to date? Based on his arsenal and arm angle?
I only brought up Perez’s splits to point out the lack of evidence (indeed contrary evidence) suggesting he’s any better vs. LHP than his meager overall stat line would indicate. I also don’t put any stock in reverse platoon splits, particularly over that small a sample, and didn’t mean for anyone to take from my comments that he should be expected to continue to be a worse hitter vs. LHP than RHP going forward. Just pointing out that while some hitters do demonstrate very strong platoon advantage from early on in their careers, Perez is clearly not one of them.
BTW, those pitcher/batter expected K rate charts change completely depending on the average K rates for batters and pitchers. They are not absolute. For example, if the average K rate for hitters and batters were 20%, then when a 20% batter and pitchers face one another, obviously the result should be 20%. But if the average were 15%, then the matchup might yield a 25% K rate. It is basically an odds ratio of log5 method of computing matchup results. One of the terms is the average rate.
I figured since it was from last year it was fine.