Hughes' postseason struggles

A few weeks ago, I used PITCHf/x to look at Phil Hughes’ interesting 2009 season. I was asked to do a follow-up on Hughes’ less-than-stellar postseason. I’d like to share a few of the things I found.

Obviously, Phil didn’t get the kind of results in the postseason as he did in the regular season:

Regular season: 1.40 ERA, 1.83 FIP, 11.4 K/9

Postseason: 8.53 ERA, 4.83 FIP, 9.95 K/9

What’s strange is that while he got absolutely tattooed, he still had a high strikeout rate. Here is his “stuff” (on the left is from his regular season bullpen stint, on the right is from the postseason):

Pitch	Average Speed (mph)	Max Speed	pfx_x	pfx_z	spin_dir
FF	94.64 94.32	97.3 96.3	-4.98 -4.54	10.33 10.40	205.80 203.47
FC	89.26 88.92	92.6 91.9	1.10 1.17	5.27 5.68	166.79 167.33
CU	77.49 77.53	80.4 79.1	7.27 6.25	-7.86 -9.44	42.78 33.73

Hmm. That’s not much different. It would make sense that if a pitcher all of a sudden starts getting hit hard, it would correlate with some sort of change in raw stuff, but that doesn’t seem to be the case here. There’s a slight loss in velocity, but I don’t see how that could have much of an effect overall. Below are the pitch results. Swing% is swings / total pitches; Whiff% is swinging strikes / swings; Wide Zone% is the vertical strike zone set by the PITCHf/x operators and a two-foot horizontal zone (one foot off of the center of the plate); Chase% is swings out of the wide zone; Watch% is takes inside the wide zone.

Pitch	Pitch#	Pitch%	Swing%	Whiff%	Wide Zone%	Chase%	Watch%
FF	554 87	66.3% 62.1%	52.2% 42.5%	27.0% 32.4%	56.7% 48.3%	26.0% 35.1%	31.8% 42.9%
FC	107 23	12.8% 16.4%	56.1% 70.0%	33.3% 12.5%	65.4% 65.2%	21.7% 6.3%	32.9% 20.0%
CU	166 30	19.9% 21.4%	34.9% 63.3%	19.0% 42.1%	44.0% 56.7%	25.9% 42.1%	41.1% 35.3%
	830 140	99.4% 99.3%	49.0% 51.4%	26.8% 30.6%	55.1% 52.9%	25.3% 30.6%	33.5% 36.5%

I’m still not seeing what could cause that much of a decrease in effectiveness. The cutter is the only pitch that didn’t see an increase in whiff rate from the regular season. There is a decrease in Wide Zone% --- I’ll get to that in a minute. Before I get to some thoughts on what might’ve been wrong, I’d like to show a final set of numbers comparing his regular season and his postseason. SLGCON is slugging on contact, or total bases / balls in play. Runs Saved is linear weights based on pitch outcomes (ball, strike, single, double, strikeout, etc); RSv/100 is Runs Saved per 100 pitches. Expected Runs Saved is calculated the same way as Runs Saved, except that it is based on batted ball types (groundball, flyball, line drive, pop up) instead of outcomes; xRSv/100 is the rate stat.

Pitch	Pitch#	SLGCON	RSv/100	xRSv/100
FF	554 88	.406 1.111	2.57 -3.88	1.82 -0.23
FC	107 22	.263 1.000	4.56 -1.61	2.83 -0.76
CU	166 30	.296 .429	0.61 2.50	-0.99 1.69
	830 140	.357 .842	2.40 -2.14	1.37 0.13

So it’s clear that something happened. Hughes turned from exceptional to horrible, although you wouldn’t be able to tell from his peripherals. I had a few thoughts as to what might have been Hughes’ problem in the postseason, so let’s get to those.

The first thing I’d like to look at in terms of graphs is his pitch location on hits in the postseason.

Nothing earth shattering here; it looks like he hung a few curveballs and left a few fastballs too low.

Moving on! One thing noticeable in the pitch results is that Hughes had a lower zone percentage, especially with his fastball, in the postseason. So, is it possible that he was falling behind in the count and then getting hit hard after falling into “hitters’ counts”? Let’s take a look:

Stint	FPS%	Ahead%	Behind%
Regular Season	62.4%	46.9%	14.4%
Postseason	60.6%	45.5%	12.1%

That doesn't really explain a whole lot. In addition to falling behind hitters at a similar rate as he did in the regular season, he maintained an excellent first-pitch strike percentage. So that theory is essentially debunked.

In his postseason games, I remembered seeing him “nibble” at the strike zone more than he did in the regular season, and I wondered if what I remembered seeing had any bearing on a) whether this was actually happening and b) if this had any effect on his performance (for example, his “nibbling” on the corners might’ve cost him some close calls from the umpire). I figured I could test this out by taking the average distance of each pitch from the center of the plate, and the mean that I got would be the “Nibble Score.” In order to filter out balls way out of the zone, I used the wide zone plus three extra inches on each side of the plate. Also, for here, I only used fastballs. This next graph shows the Nibble Score in a variety of different count situations, dictated by run expectancy. Negative numbers indicate a count favorable to the pitcher; positive numbers indicate a count favorable to the hitter. The count-based run expectancy is 0 on the first pitch of an at-bat.

The sample size is small for the postseason stint, so I’ll let you draw your own conclusions from the graph. There doesn’t seem to be that much of a difference between the overall nibble-ness (0.55 in the regular season, 0.58 in the postseason), in which I’m working with a larger sample size, so I’ll put that theory to rest now.

The last thing I wanted to look at (for now) was Hughes’ release point by pitch. It’s pretty clear that the curveball is coming from a higher slot, and it makes me wonder if hitters were able to sit on his fastball after seeing the lower release point. But still, the curveball is coming from a bit higher in the regular season plot, and he obviously had great success. For what it's worth, here are the charts.

So, I couldn’t find any statistic that gave any clear answer as to why Hughes struggled so mightily during the postseason. He was getting ahead on hitters, he had good velocity, and he had the ability to miss bats. Because of this, I’m not overly concerned about Hughes. However, I looked at some video of the postseason earlier today, and in the (small) sample that I looked at, it seemed as if Hughes was missing location within the zone --- Jorge Posada had to move his glove a lot. This isn’t quantifiable by any metrics we have now, but that could be the most important thing --- pitching to the situation. Although we can use PITCHf/x to give us a good indication of what happened, it can’t tell us what the pitcher intended to have happen, and I think that in this case, this was what burned Hughes the most.

Gameday PITCHf/x data is from MLB Advanced Media; it can be easily accessed here and here. Other statistics are from Fangraphs.