For Hughes: Regression, or Improvement?

(Photo by Nick Laham/Getty Images)

Phil Hughes is having a tremendous month of May. The 25-year-old, who left April a weak link in a weakening New York rotation, is 3-2 with a 3.45 ERA in five starts over the past three weeks. Hughes has a 4.1 strikeout to walk rate over the past month and batters are hitting just .230 off the righty, giving him a WHIP of 1.12 since May first. That’s a staggering three quarters of a runner fewer per innings in May than in April. Hughes has pitched well enough to avoid demotion upon the return of Andy Pettitte, and has strung together his five best Game Scores of the season in his last five starts, including a six inning, two run, seven strikeout performance against the Royals last night at home.

Despite a strong spring training, and all the optimism in the world heading into the regular season, Hughes’ has gotten off to a rough start. Even over the past month, Phil’s FIP (4.37) remains below league average and his flyball and home run rates remain far too high. His ERA is much worse even than his FIP. In medium and high leverage situations, of which he has faced 96 on the season, Hughes’ has surrendered a wOBA of .385. Derek Jeter has a .377 wOBA on the season. Regardless, his last five starts have been miles better than his first four. He’s going deeper into games, throwing far more strikes, surrendering fewer hits and walks, surrendering fewer home runs, and has an ERA more than four runs lower than his ERA in the month of April.

We are all happy to see Hughes succeed now. The question is whether the adjustments Hughes has made are sustainable. Are we simply seeing the inevitable BABIP and HR/FB correction? Or is there something more here?

Hughes' BABIP after each start, 2010-2012.

The former is eminently possible. Hughes suffered through a .373 BABIP in the month of April. His BABIP in May? A measly .253. His HR/FB rate is still high, but fell from 17.9% to 12.8%, a significant drop. After stranding just over half the batters who reached base against him in April, Hughes has stranded nearly 90% of such batters in May. For the season, Hughes’ “luck” metrics are now within reasonable range. His .297 BABIP is above his career average but in line with Major League average. His strand rate, 72%, is also fairly normal. His HR/FB rate is still quite high, at 14.7%, and that largely explains the difference between his ERA/FIP and xFIP. Yet as the dust settles on two highly volatile months, Hughes’ line is believable, and on the season could look much like his line over the past year and a half.

That would be an accomplishment after how Phil started the season. For all his ups and downs, the Phil Hughes of 2010 was a decent Major League starting pitcher. I think we all believed, at some level, that behind the guy we were watching in April, somewhere, lay the Phil Hughes of 2010, though probably not the Hughes of 2006 or 2007. As his luck has corrected, that pitcher has emerged, volatility intact. Certainly another season like 2010 would be a blessing for the Yankees. It would not belly criticism of Hughes, it would not undermine any “Hughes to the bullpen” case necessarily, but it would provide a degree of value in a difficult time for the team’s pitching.

Still, just as we suffered through a statistically unbelievable degree of failure from Phil in the first month of the season, it’s difficult not to try and draw conclusions from Phil’s success in the second month of the season. Maybe he’s doing something worthy of that 3.45 ERA. Maybe this isn’t just normal volatility and regression. It’s only five starts but Hughes looks different. He has made an adjustment. After throwing his cutter over 12% of the time in April, Hughes has essentially abandoned the pitch in May, throwing more fastballs and breaking balls instead. Our own Eric Schultz documented this shift back on May 8th, after Hughes’ second start of the season without a cutter, his second start of the season pitching at least two outs into the sixth inning.

Phil's BB/9 after each start this season. Arrow indicates adjustment to pitch selection.

Hughes’ cutter had been especially poor in the early going. According to Fangraphs, the pitch has been worth -6.7 runs per 100 throws this season. By abandoning the pitch, establishing his fastball, and relying more heavily on his breaking pitches, Hughes has seen better results. If we believe that the cutter had something to do with Hughes’ high BABIP and HR/FB rate in the first month of the season (that RAR per 100 figure certainly suggests it did), it’s abandonment could well be a catalyst for their regression to better than league average figures. Hughes has also been able to limit his walks, by more than a full walk per nine innings, crucial for a pitcher who’s OPS is 59 points higher with runners on base and who surrenders far more than a home run per nine innings. Perhaps most importantly, Hughes has been more efficient. Before May 1st, Hughes averaged four innings a start. Since then, he’s averaged six and a third innings a start.

I will suggest that Hughes’ recent performance is undoubtably, partially the result of a regression to the mean of his BABIP, HR/FB rate, and other “luck” statistics. Neither his early season, nor recent, BABIP and strand rate figures are within a reasonable range for any but the most extreme of pitchers. Yet the timing of this transition, coinciding with Hughes’ adjusting his pitch selection for a greater degree of control, suggests also that there real improvement here. We aren’t simply seeing the inevitable improvement over the Phil Hughes of April, but a better Hughes than we’ve seen in a while, a Hughes capable of staying in the starting rotation for now should he not fall back on old habits.