I can’t figure out why Shane Greene was so bad in 2016. Among AL pitchers with at least 40 innings thrown last season, Greene’s 5.82 ERA was the 19th-highest. Pitchers above him include Mark Lowe and Anibal Sanchez, two players who we know were bad last year. The rest of the list isn’t all that surprising, aside from whatever happened to Oakland Athletics ace Sonny Gray. According to Baseball Reference, Greene was -0.6 wins below replacement level last season. Even if you exclude his first four starts, Greene still managed a 5.67 ERA as a reliever.
However, every other metric in existence suggests Greene was miles better than his ERA showed. His 3.13 FIP was the 27th-best in the AL, just a couple spots behind Yu Darvish. Greene also managed a 3.95 xFIP, though one might argue that measure doesn’t account for how well he should be able to suppress home runs. Other more advanced formulas like SIERA and Deserved Run Average (DRA) peg Greene in the high 3s. This isn’t elite by any means, but we’re talking about a two-run discrepancy here.
Let’s dive deeper. Greene’s strikeout and walk rates weren’t amazing, but still better than average. He managed a 12.7 percent swinging strike rate, which was well above the AL average. He induced more soft contact than all but eight AL pitchers, and was just outside the top 10 in hard contact allowed. His ground ball rate was just above 50 percent, and his pop-up rate was around league average. He only gave up three home runs all year. Alex Chamberlain’s xBABIP (expected batting average on balls in play) calculator says that Greene’s .327 BABIP was 31 points too high.
That still doesn’t scratch the surface, though. If you apply Greene’s .296 xBABIP to his 2016 numbers, you shave off a mere five hits allowed in 60 1⁄3 innings. Even Statcast thinks he was pretty good; he ranks well better than average in all types of exit velocity allowed, and just misses being screenshot-able from the top of the list. He locked down with runners in scoring position, holding opponents to a .217 batting average. He even stranded all 20 baserunners he inherited last season. This would not affect his ERA, but is impressive nonetheless.
The dog days of summer
A quick glance at Greene’s game logs show that his ballooned ERA is largely the result of a bad month-plus stretch. Entering August, Greene had a 4.14 ERA. If you exclude his four April starts from that number, he had a 2.78 ERA as a reliever in two months of work. His 24 strikeouts and five walks in 22 2⁄3 innings of work suggest it wasn’t a fluke, either.
Then August happened. He gave up two runs on three hits (including his first home run of the season) on August 2, and it all spiraled out of control from there. From August 1 to September 7, Greene allowed 16 runs in 14 2⁄3 innings, a 9.82 ERA. He gave up all three home runs of the season during that stretch, accounting for six of the 16 runs. His ERA as a reliever increased to 5.54, nearly three runs higher than it had been entering the month.
There was little rhyme or reason as to why Greene fell so far. His fastball velocity held steady during that month, and we did not hear any news of injury. His pitch usage was largely the same as well. His strikeout rate dipped from the high 20s to 23.1 percent in August and 21.8 percent in September, but held steady on a per-inning basis. We have played the “well, if you remove this outing” game before — hey there, Rick Porcello — and it drops his ERA considerably, but that’s how things play out with relievers.
Let’s get ahead of ourselves
There isn’t much of a correlation here, but Greene’s success on the first pitch of an at-bat might give us a clue as to what happened.
|Month||First Pitch Strike %||Batting Average Against|
|Month||First Pitch Strike %||Batting Average Against|
We’re dealing with small samples here, but the data we do have suggests that Greene’s success largely depended on getting strike one. This is true with all pitchers — league-wide OPS varied by over 200 points depending on whether the first pitch was a strike or ball — but it was especially true with Greene. When he got ahead 0-1, Greene limited opponents to hitting just .162/.208/.222, a .430 OPS. If hitters went ahead 1-0, they fared much better, batting .341/.449/.489. That’s a 507-point difference in OPS!
Why was this the case? It appears Greene became very predictable if he fell behind in counts.
Greene went to some form of fastball roughly 90 percent of the time when he fell behind last season, including 98 (!) percent of the time against left-handed hitters. His cutter has good velocity separation from his fastball, but isn’t the type of cutter that induces a lot of weak ground balls. He averaged a 46 percent ground ball rate on the cutter last season, which isn’t all that high. He also had trouble locating the cutter at times against left-handed hitters, failing to bury it down and in.
Opponents still fared better against his fastballs than his cutter and slider, but all three home runs Greene allowed came off the cutter. He may want to get a little less predictable when falling behind, but tightening his command will be just as important.
Where do we go from here?
That’s the question, isn’t it? Greene’s ERA indicated that he had a poor season in 2016, but everything else, save for that one wonky month, suggests that he was much better than that. We don’t know what caused his decline in the second half — fatigue may be a factor, but his velocity didn’t dip at all — and probably won’t know which Shane Greene to expect until well into 2017. If he can hone his command a bit more, the Tigers may see more of the reliever that dominated the American League in June and July. If not, then we will be asking the same question again for yet another winter.