Cahill’s adjustmentsby Lucas Apostoleris
April 28, 2011
After winning 18 games as a 22 year old last year, Trevor Cahill is off to a nice start again this season as he's 3-0 in his five starts with a 2.30 ERA.
So far this season, his strikeouts (23 percent per batter, up from 15 percent last year) and whiff rate (.202 per swing from .164) are up, and his control is better (six percent walks and hit batsmen per batter, down from eight percent), which has more than compensated for a slight dip in groundball rate (down about 10 percentage points from last year).
Expecting to find something startling different in either his pitch selection or results in 2011 compared to the previous two years, I classified all of Cahill's regular season PITCHf/x games and began to look for potential differences.
The first thing to observe in Cahill's game is how he mixes his pitches. Cahill is a sinkerball pitcher, throwing an 86-92 mph sinking two-seam fastball over 55 percent of the time in his career. The pitch gets a ground ball about 57 percent of the time it is put in play, which is substantially above average for two-seamers. He'll also throw some four-seam fastballs that are harder (91 mph) and have less movement.
He'll throw a changeup, typically around 80 mph, to both left- and right-handed hitters. His other two pitches, a curveball (high 70s) and slider (low 80s), have been at the center of the most adjustment in his young career. Below are two graphs that show Cahill's pitch mix in each major league start. The one on the left is against left-handed batters, and the one on the right is against right-handed batters.
When he first came up, Cahill used his slider a lot—mostly against righties, but plenty to lefties as well. The curve was a rare bird in 2009, representing less than three percent of his total pitches. At the start of the 2010 season, you can see that Cahill abandoned (only five in all of 2010) his slider against lefties but kept throwing it against righties.
At the same time, his curveball started to become an important piece in his repertoire—for a time from August to September of 2009, he appeared to ditch the curve completely (especially in '09, his curve and slider would blend together a little bit, so there is some subjectivity in the pitch IDs), but his reliance on it increased in 2010. Throughout the season he used both, but as you can see from the graph, he typically wouldn't mix them both evenly within a start. Sometimes he would be more curve-heavy, other times more slider-heavy.
Finally, you may notice that there's no more "green" on the last few bars. Cahill has not thrown a single slider all year long. As a result, he's throwing a lot more curveballs as his out-pitch against righties.
So, other than the missing slider, what's changed about Cahill? I looked at some performance metrics on Cahill this year compared to the previous two years, and most things look pretty much similar.
Whiff rates (swings and misses per swing) and watch rates (takes on pitches inside the strike zone) are two of my preferred tools for looking at a pitcher's ability to get strikes. The table below shows those and a few of my other favorite metrics for Cahill's pitches in 2009-2010.
By the way, I'm showing two new ones today - pERA and pxERA. It's just the run values I usually work with, except they're formatted to look like ERAs - pERA (pitch ERA) is from outcome-based run values and pxERA (pitch expected-ERA) is from batted-balled type-based run values. It's not exact and probably doesn't add a whole lot, but I wanted to try it out and see if it's more intuitive than straight run values.
|#||Swing Rate||Whiff Rate||Zone Rate||Watch Rate||Ball Rate||GB Rate||pERA||pxERA|
As you can see, Cahill's isn't a superior bat-misser, but at 35 percent, the curveball whiff rate is quite good. The changeup isn't bad either (compare here). The slider, however, was not very effective, as it had both a low whiff rate and ground ball rate. It showed improvement from 2009 to 2010, increasing the whiff rate from 23 percent to 30 percent, but it was still reasonable for Cahill to show more confidence in his change and curve.
The whiff rate on Cahill's curve improved substantially as well from 18 percent in 2009's limited sample to just under 39 percent in 2010. So, since the curve and change are more effective than the slider, the lack of sliders is what has made Cahill better this year, right? I would've thought so, but he's actually getting fewer whiffs and zone-takes on his offspeed stuff in 2011:
|#||Swing Rate||Whiff Rate||Zone Rate||Watch Rate||Ball Rate|
The difference in whiff rate is mainly coming from a few extra empty swings on his sinker than he typically gets, and not from the extra curves in place of the sliders. While it would've made a nice and logical explanation for Cahill's strikeout spike, it doesn't appear to actually be the case. On the other hand, I think it's reasonable to assume that the curveball whiff rate improves towards its 2010 level.
What should we expect from Cahill moving forward?
Cahill is a very good pitcher based on his ability to get ground balls and maintain a good walk rate (it's improved since his rookie year). I do think that despite his slider's improvement from 2009 to 2010, the curve is still a better option as a strikeout and groundball pitch in the future.
As for what Cahill's doing right now (K/9 rate over 8.0, well over where he was at last year, but with slightly lower whiff rates on his offspeed pitches), I'd highly doubt that his strikeout rate stays that high; however, given his decreased slider usage and increased curveball usage, I would expect him to pick up more strikeouts this year than last.
Being a groundball pitcher who has exhibited good control, picking up strikeouts certainly isn't Cahill's first priority, but it can't hurt to be able to throw a strikeout pitch when you want to. By showing an increased reliance on his curveball, Cahill is putting himself in a better position to get more strikeouts and thus improve a big part of his game.
References and Resources
Gameday PITCHf/x data are from MLBAM and are used here courtesy of Joe Lefkowitz's PITCHf/x tool. Pitches have been reclassified by the author.