Philip Humber: Not A Fluke
Over the weekend, Philip Humber had the game of his life, throwing the 21st perfect game in baseball history against the Seattle Mariners. While the Mariners have a lousy offense and Safeco Field is a fantastic place to pitch, those factors shouldn’t diminish what Humber accomplished. A lot of good pitchers have faced a lot of lousy offenses over the years, and only 20 men before Humber had managed to go 27 up, 27 down. This is the apex of a single game performance in the sport, and Humber has now etched his name into the history books.
He’s also serving notice that last year’s breakout season may not have been a fluke.
Before last year, Humber was pretty well traveled for a relative youngster. The third overall pick in the 2004 draft by the Mets, he never developed into what they had expected, and he ended up being part of the package that New York shipped to Minnesota in exchange for Johan Santana in 2008. The Twins kept him around for two years – though they did pass him through waivers in April of 2009, where he went unclaimed – before he signed a minor league contract with the Royals. After spending most of 2010 with the Royals Triple-A club in Omaha, the A’s claimed him off waivers right before Christmas, only to designate him for assignment a few weeks later when they needed the roster spot. This time, the White Sox grabbed him, and teaming up with Don Cooper in Chicago has been the best thing that ever happened to his career.
Thanks to Jake Peavy’s shoulder problems, Humber ended up snagging the #5 spot in the White Sox rotation last spring, where he continued with his usual pitch-to-contact approach, throwing strikes and hoping opposing batters hit the ball right at his teammates. It worked pretty well last April, as his .212 BABIP kept runs off the board, and his good results caused the White Sox to go to a six man rotation in order to keep giving Humber starts. It seemed like an experiment that wouldn’t last, though, as Humber’s 4.21 xFIP in April and 4.40 xFIP in May suggested that he was still just a back-end starter at best, and his results would begin to trend downwards sooner or later.
It took a little longer than you might have expected, but eventually, the luck dragon did indeed fly away. Humber posted BABIPs of .400, .333, and .313 in the final three months of the season, and his second half ERA of 5.01 gave off the impression that Humber had just reverted back into waiver-bait status. What it hid, however, is that Humber was actually getting better.
When Humber got to Chicago, he threw what he referred to as a cutter, even though Pitch F/x called it a slider. Cooper didn’t like the pitch, and instead suggested Humber move to a more traditional slider with harder breaking action. The differences are pretty striking.
Here’s Humber’s game chart from his April 9th, 2011 start against Tampa Bay.
You can see those cutters clustered near the 0 horizontal movement axis, hanging out in that 86-88 MPH range. He didn’t throw a lot of them, and pitched primarily off his fastball/curve/change repertoire, none of which are really legitimate out pitches. After taking Cooper’s advice, however, he began throwing a more traditional slider. Here’s his game chart from his July 17th start against Detroit.
While you can still see a couple of pitches that look more like his previous cutter, there’s a much larger cluster of pitches around the 0 horizontal movement axis with velocity bunched in the 84-86 range. This is a pitch Humber just wasn’t throwing before, and it gave hitters a new wrinkle to deal with. In this particular start, Humber struck out eight of the 24 batters he faced – he never struck out more than five batters in any of his starts in April or May.
As he refined the slider, his strikeout rate shot up dramatically. Here’s Humber’s K% by month:
April: 15.9%
May: 12.3%
June: 18.3%
July: 19.0%
August: 14.3% (14 IP, DL stint)
September: 22.6%
Even in Triple-A, Humber had never managed strikeout rates in line with what he posted down the stretch last year. With the addition of a harder slider in lieu of his cut fastball, Humber now had the ability to generate swinging strikes. The results down the stretch didn’t look great, but Humber’s ability to miss bats looked like a promising development, especially considering Cooper’s long standing reputation for turning mediocre pitchers into quality arms.
Now, obviously, we’re only two starts into 2012, but here’s Humber’s game chart from Saturday:
The first thing that jumps out is the big cluster of sliders right in the middle – this is a pitch that he’s become very comfortable with, and is the pitch that he called upon to get Brendan Ryan to chase for the final out of the game on Saturday. But, beyond just the move towards more sliders, look also at where his fastball and change-up velocity are – both were up significantly over last year, and his fastball actually got near 95 a couple of times.
When he got to Chicago, Humber was a guy who threw a 90-92 MPH fastball and an 86-88 MPH cutter. In his first two starts of 2012, he’s been a guy throwing a 90-94 MPH fastball and an 83-86 MPH slider. The differences in velocity and movement have given him distinct swing-and-miss weapons, and when combined with his change-up and curveball give him four pitches he can throw for strikes.
We’re only dealing with a 2012 sample of 211 pitches, but it’s worth noting that opposing batters have only made contact with 66.7% of the pitches they’ve swung at in his first two starts. That mark ties with him with Daniel Bard for the lowest contact rate of any starting pitcher in baseball so far. This isn’t just Humber getting balls to go right at opposing fielders – this is Humber throwing strikes and missing bats.
Throwing a perfect game doesn’t make Humber an ace, and neither does shutting down the Mariners offense in Seattle. However, Cooper’s adjustments to his repertoire began showing signs of effectiveness last summer, and he’s carried over those changes to the start of the 2012 season. He’s always had quality command, but now it looks like he may have developed a legitimate out-pitch as well. If he can keep getting hitters to chase his slider, Humber could end up as the latest in the long line of Don Cooper success stories.
Dave is the Managing Editor of FanGraphs.
Dave,
Is there an MPH typo in here, or am I misreading the chart?
“You can see those cutters clustered near the 0 horizontal movement axis, hanging out in that 86-88 MPH range.”
Looks to me like the cutters in the first charts are in the 78-83 range, which is also more in line with your argument.
I think you’re looking at the chart below that section as opposed to the one above it. He’s referring to the chart above that section of text.
Nope. I’m wrong. You’re looking at the CU — that’s change up. The SL — slider — is what Dave’s calling the cutter.
Wrong again. CU is curve ball, not change up. Still, that’s what you’re looking at.
Boy am I schizophrenic today, aren’t we?
Gotcha, thanks. I thought CU was cutter. Still, not too many slider/cutters in that area on the first graph, I only see two.
The Pitch F/x classification algorithm isn’t always so great at identifying the differences between fastball types or cutter/sliders, so it’s generally better to just look at clusters in velocity/movement/spin rather than relying on the algorithm to tell you what pitch is what.
I may be wrong, but I believe “FA” is a catch-all fastball categorization since it can be very difficult to tell the difference between two seamers, four seamers, and cutters just from computer analysis. Where do you draw the line? The pitches that Dave refers to – based on speed and horizontal movement alone – are likely cutters. It appears that they are sometimes miscategorized as sinkers (SI) due to the drop in the pitch.