Six months ago, we were all idiots. We were all idiots because we figured the Yankees were idiots for letting Masahiro Tanaka pitch. It felt like elbow surgery was inevitable, UCL tears being things you don’t just play through, so it felt strange to see the Yankees in denial. Oh, we all thought we knew what was better. Turns out the Yankees might’ve been on to something. Turns out those doctors have more than just pieces of paper. Tommy John might still be an inevitability. Greg Holland pitched through something like this for a year. The elbow still got him. The elbow might still get Tanaka, but tonight he’s starting a playoff game. Tonight, he’s the guy, at least until Andrew Miller and Dellin Betances are the guys. It’s safe to say now it worked out.
Remain in the flashback, though. Go back six months to when Tanaka made his first start. There was so much attention on his fastball — on its usage, and on its velocity. Observers thought the fastball would be a dead giveaway. They thought the injury would make the fastball worse, and they thought Tanaka wouldn’t be able to cut it without a decent fastball. Just about everyone, after all, builds off the fastball. It’s the pitch at the center of the pitch-type solar system.
In the end, Tanaka’s fastball actually gained velocity this season. He started 24 times, with a mid-3s ERA, and his average fastball was up the better part of one mile per hour. In a sense, the fastball wasn’t a giveaway. But if you stare long enough, you see two things. One: overall, Tanaka was pretty good. Two: Tanaka’s fastball was pretty bad.
You’re familiar with our pitch-type values, I assume. They should be used with care, and you shouldn’t make much of small differences, but the numbers, in general, do a good job of separating good pitches from bad ones. A year ago, in Tanaka’s debut season, his fastball was 10 runs worse than average. It was worth -1.25 runs per 100 fastballs. Right there, we already had evidence that Tanaka could get by with a sub-par heater. Now look at this season. His fastball, this time around, was 18 runs worse than average. It was worth -2.35 runs per 100 fastballs. Presumably, you can recognize that as bad. It’s helpful to put it in context.
We have pitch-type information stretching back to 2002. So I gathered that, looking for pitcher seasons with at least 100 innings. For each pitcher, I added together all the run values to get an overall run value, above or below average. Then I plotted that with run value per 100 fastballs. It can provide a little indication of how important it has or hasn’t been to possess an effective heater.
You see 2015 Masahiro Tanaka highlighted, and he’s pretty well removed from the pack. There’s a pretty clear linear relationship overall, as you’d expect, but Tanaka defies that. Out of the pool of 1,998 pitcher seasons, this year’s Tanaka ranks eighth from the bottom in average fastball value. And he’s the only pitcher in the area to have still been overall above-average. To put it another way: 1,026 pitchers in the sample finished with an overall run value of at least +0.1. Out of those guys, Tanaka had the least-valuable fastball, at -2.35 runs per 100. Next-worst is 2007 Sean Marshall, at -2.06. Next-worst from just this season is Jason Hammel, at -1.10. Tanaka’s mark was more than a full run worse.
About that sample of 1,026 — last year’s Tanaka is sixth from the bottom in average fastball value. So this season just took last season and turned it up. Overall, Tanaka finished five runs worse. His fastball alone was almost eight runs worse. He now has a total sample of 44 regular-season starts, and it’s pretty evident he can be good with a fastball that hurts him more than it helps him.
Neither the Yankees, nor Tanaka, are idiots. This year, he trimmed his fastball rate from 41% to 33%. In the season’s final month, he came in at 26%, throwing fewer fastballs than splitters. Tanaka needs to have a fastball, but it isn’t a weapon for him on its own. Its only a weapon in how it might leave hitters unprepared for the other pitches. According to PITCHf/x, both Tanaka’s four-seamer and two-seamer this year yielded slugging percentages well north of .600.
Out of curiosity, I looked for pitchers with similar repertoires, based on pitch types and usage. The best comp I found for 2015 Tanaka was 2012 Mark Buehrle. Then, of course, there was 2014 Tanaka, followed by 2006 Odalis Perez, and 2011 Freddy Garcia. 2014 Hiroki Kuroda is tied with Garcia. Hisashi Iwakuma isn’t far away. Neither is Jamie Moyer. (I combined changeups and splitters for this exercise.) Tanaka has adopted an uncommon approach, shifting even more away from fastballs than before, and the list of comps runs dry pretty quick. Right now he’s someone who’d be classically described as “crafty.” That’s not a label we would’ve expected when Tanaka first signed with New York.
If you think about how Tanaka works, he’s most vulnerable when he needs to throw strikes. When he can put his splitter somewhere out of the zone, the hitter’s out of luck. But this year, after Tanaka fell behind 1-and-0, he allowed baseball’s eighth-worst slugging percentage. In the combined hitter-friendly counts, Tanaka again allowed baseball’s eighth-worst slugging percentage. When the hitter is in charge, he can eliminate certain pitches, and Tanaka has to think more about the zone. This much is true for everyone, but it can be particularly bad for a pitcher whose greatest strength is getting guys to chase a pitch diving for the dirt. From Baseball Savant, here’s where Tanaka located when behind in the count:
Meanwhile, here’s where he located when ahead:
The distribution is as you’d expect, and while Tanaka still made the occasional mistake when in front of the hitters, they didn’t happen as frequently. Tanaka is most vulnerable when the hitter assumes he needs to throw a fastball for a strike. The fastballs just aren’t good enough to overpower the opposition. And what Tanaka really wants the opposition to do is chase, which it’s less willing to do when it can afford to be patient.
How much this is going to mean for tonight, I don’t know. Trying to forecast one baseball game is like trying to forecast the flight path of a bug that you’re trying to swat. The one time Tanaka faced the Astros this year, he got hit around a bit, but then that game was in Houston, and this swing somehow turned into a dinger:
Sometimes that’s how baseball happens. A game is a lot of little individual baseball components, and they stack together all baseball-y, and in the end you have something you recognize as baseball even if you couldn’t have known what shape it would take. You know how this works. There are probabilities, and then the test is run one time, and we all get to try to interpret it after. It’s a weird thing we do, and it’s impossible to know how much of a role Masahiro Tanaka is about to play in the outcome of the wild-card contest.
But if you leave this knowing one thing, know this: of all the above-average pitchers from at least the last 14 years, Masahiro Tanaka just threw the least-valuable average fastball. And they said it couldn’t be done.
Jeff made Lookout Landing a thing, but he does not still write there about the Mariners. He does write here, sometimes about the Mariners, but usually not.