Both Sides of the Phil Hughes Argument

So much of the analysis we do in baseball follows a pretty simple formula. We notice a player is getting different results, we look for explanations for the change, and then we speculate about whether the latter caused the former and if it will continue. Baseball includes so much random variation and you have to make decisions all the time about how much information is enough to rule out randomness as an answer. When it came to Phil Hughes in 2014, we didn’t have to worry too much about randomness as an explanation.

Hughes rattled off the best K/BB ratio of all time and was somewhere between very good and excellent for the first time in his career. While we can chalk lots of leaps up to random variation, Hughes seemed immune because he was such a drastically different pitcher in Minnesota compared to New York. It appeared to be quite clear why Hughes was so much better. He stopped issuing walks and pounded the strike zone. He bumped up his strikeouts as a result, and seemingly cut his home runs because he was in better counts and not in Yankee Stadium.

Even with bad defense and an above-average BABIP, Hughes thrived. FIP liked him a little better than ERA, but even by a basic RA9-WAR, he was a 3.5 to 4.0 WAR pitcher. The question we were all asking ourselves was if Hughes could keep it up. We didn’t wonder if he was just getting lucky. He pitched differently and the results got better.

Year	TBF	K%	BB%	HR/9	BABIP	ERA-	FIP-
2010-2013	2521	19.0%	6.9%	1.42	.295	111	105
2014	855	21.8%	1.9%	0.69	.324	92	71
2015	326	14.7%	2.5%	1.53	.313	125	115

Until they didn’t. A year later, Hughes is having arguably his worst season since settling in as a full-time starter. If nothing else, he’s fallen back to the pre-2014 version of himself when it comes to the outcome variables. Naturally, one would think this occurred because Hughes was unable to keep doing the things he was doing that made him great in 2014. But that’s not really true. Hughes looks very much like the same pitcher. He’s pounding the zone with regularity and his pitch mix has changed only slightly.

Hughes is throwing basically the same pitches to basically the same locations and he’s getting clobbered. How can something that worked so well in 2014 work so poorly in 2015?

The shortest answer is contact rate. In 2014, Hughes threw a ton of strikes, batters swung a lot more, and didn’t make much additional contact compared to previous years. That’s a great formula. In 2015, he’s still throwing strikes and getting swings, but the contact rate has shot up.

Year	Zone%	Swings%	Contact%
2010-2013	53.0%	48.8%	82.7%
2014	61.1%	56.6%	84.2%
2015	58.7%	53.6%	90.0%

This is a story about batters adjusting to Hughes’ new way of doing business. In 2014, his strike-throwing mania was new. Hitters saw more strikes and started to swing at them, but they weren’t getting the bat to the ball much more efficiently than before. After a year of seeing Hughes fill up the zone, they’re not missing when he attacks them.

The whole thing seemed like a very dangerous gambit in the first place. August Fagerstrom wrote last November that Hughes’ transformation was similar to Cliff Lee’s, but the caveat was there at the time: Lee’s transformation was incredibly unusual and it also required that he learn to strike batters out. The odds, Fagerstrom told us, were heavily stacked against Hughes ascending into Lee’s category.

The feeling, I would argue, after Hughes’ 2014 season was that he had become a better pitcher, but not one that would hover around the Cy Young race. The Twins extended his contract and he looked to be a nice building block for a team that was assembling a nice collection of prospects. We rightly expected some regression from Hughes. Our depth charts listed his projection at 3.66 ERA and 3.21 FIP. He was slated to maintain the strikeouts and regress on the walks and home runs. But overall, we pegged Hughes as a four-win pitcher over 205 innings. A slightly worse ERA and a moderately worse FIP. That seemed entirely appropriate.

Instead, he’s allowed 13 HR in 12 starts, including at least one in nine starts and two in four starts. The new Hughes was supposed to be better at home-run suppression than the old Hughes, but not on par with his 2014 version. Your first instinct is to pump the brakes and acknowledge that something as temperamental as HR/FB% could easily be shining an unfair light on Hughes over one-third of the season. That’s true, but his 2014 HR/FB% looks like the outlier in the context of his career. At the time, you might suggest it had to do with his park and newfound strike-zone wizardry, but after watching him in 2015, it’s looking more like good fortune or unprepared hitters drove 2014.

It’s also not a question of sequencing. Here are his wOBAs against as a starter over the last four seasons:

• 2012: .329
• 2013: .354
• 2014: .295
• 2015: .347

If anything, his wOBA indicates he should actually have allowed an extra run or two this year, but it’s well within the margin of error. In other words, Hughes is doing the same thing he did in 2014, but the results have been much worse. He’s allowed 13 home runs in 2015, but his 2014 HR/9 would only have gotten him to six through this many innings. Just for fun, let’s take away seven home runs from his 2015 and generate a FIP and then let’s take seven home runs away from his 2015 and generate a new wOBA and expected runs allowed.

Year	FIP	wOBA	RA9
2015	4.41	.347	4.81
Minus 7 HR	3.23	.300	3.84

This is very much a question of baselines. If you didn’t know about 2014, you’d have no difficulty believing that the Hughes you’re watching now is the real Hughes. But if you’d only seen 2014, you’d have no trouble believing his 2015 struggles are entirely about his HR/FB% and we all know how fluid that can be this early. The 2014 seasons is probably more relevant than any one prior season, but it’s also probably not more relevant than his entire history either.

It would be very easy for us to look at his numbers this year and proclaim, “Aha! I knew he wasn’t really that good!” There’s a strong temptation do that when a player returns to being someone you thought they were. It’s an “I told you so” moment of the highest order.

At a fundamental level, the difference in result from 2014 to 2015 for Hughes has been home runs. If we could really understand the degree to which they were flukes versus legitimate, it might shed some light on this whole thing. Hughes has became a different pitcher over the last two years, but the results only stuck around for part of the time. There are two explanations. The first is that he’s getting unlucky now and the second is that hitters have learned to handle the new Hughes much better.

Hughes’ home runs have all been thrown to a home-run-happy part of the zone, so it’s not like we can chalk it up to Miguel Cabrera hitting one twenty three inches off the plate or something.  But here’s the thing. I watched each one of these, and setting aside the one where the camera misses the setup, he misses his spot in every single instance except one. And this one doesn’t really seem like his fault. Nelson Cruz was largely to blame.

In the other 11 cases (all fastballs of some kind), Hughes was trying to throw a pitch in a much different spot and it leaked back across the plate. We don’t have wholesale command statistics, so there’s no way to know if he’s missing the target more than last year or more than a normal pitcher, but it does give you some hope for Hughes that so much of the damage has come when he did something he didn’t mean to do.

It appears as if the hitters are sitting on pitches in the heart of the plate and they wouldn’t have made nearly as solid of contact if he had found the edge as he intended. Maybe that’s a cop-out answer in search of a more interesting conclusion, but he really doesn’t seem to be doing a worse job of locating his pitches overall. He’s still filling up he zone and according to Bill Petti’s Edge%, he’s not throwing pitches in the heart of the plate more often than last year.

It’s inarguable that batters are doing more damage against him. The question is if this is a situation in which they simply happen to be doing an inordinate amount of damage against his mistakes or if they have actually adjusted to the new Hughes and this will continue. If you look at the swing rate against his pitches in the center of the zone (Zone 5 on Baseball Savant), it’s 75% in 2014 and 74% in 2015. The contact rate on those pitches drops from 90% to 89% from 2014 to 2015. Here are the batted balls:

The home runs are the problem, obviously, but nothing else looks too out of place. It really comes down to the contact that’s being made in the heart of the zone being more deadly this year. The home runs are happening on misses over the plate. Phil Hughes is almost certainly not the ace his numbers suggested last year, but his 2015 might also include bad fortune in the other direction.

That’s actually pretty incredible if you think about it. Nearly everything that could possibly break in the Twins’ favor is breaking in their favor. They’re 33-26 despite not hitting or pitching very well as a unit. If that was enough against the odds, the team is also having a terrific stretch without any success from the staff’s leader.

Should they expect to see him perform like he did in 2014 again? Probably not. Yet there are some indications that he hasn’t totally regressed to the pre-reinvention arm we saw in New York. If the Twins are going to make any kind of run, they’ll need more from Hughes. Despite his early season struggles, it’s not out of the question that they’ll get it.