Warthen believes Santana may have unintentionally changed the grip on his four-seam fastball, causing it to cut rather than stay true.
“It’s going to decrease the velocity if it’s cutting like it is,” Warthen said. “The two-seamer is cutting at times also. Physically, he’s in good shape. And the arm feels good. We just have to find out what the delivery issue is right now.”
The rub of PITCHf/x
PITCHf/x jumps out as the tool of choice to check on this. First, we have to deal with two significant obstacles:
- Pitch classification
- Gameday’s pitch identification system is impressive, but flawed. It can have a hard time discerning two- and four-seam fastballs. Heck, we all do. But the hand care and feeding of the pitch IDs is an important, if not necessary, step if we want to tease out small changes in movement and speed. This won’t be discussed further in this article, but I am using my own classifications.
- Park-to-park differences in PITCHf/x
- Park effects, and calibration issues, have been quite the topic of discussion lately. Mike Fast, Tom Tango and I have recently discussed the issues with the PITCHf/x data from Yankee Stadium—both versions of the park. I’ll make a halfhearted attempt to deal with them here, if for no other reason than to show how tricky a problem it can be.
Let’s look at the raw data from Santana. “Raw” implies I’ll clean up the data later. Which is true, although “later” refers to “some other time.” Consider this an introduction and case study, both of which will be continued.
Johan Santana 2009
Starting with a survey of Santana’s 2009 starts, one can see how quickly the raw data seem to confirm Warthen’s assertion. Each of these charts includes the average (with one standard deviation error bars) and the max/min values for three measures: speed at release (55 feet from home plate), horizontal spin movement and vertical spin movement (aka rise and sink).
Click images for a larger version in a pop-up window
First, speed. Looks like a slight, recent downward trend.
Horizontal movement, which is the whopper … more than two inches of apparent difference from Santana’s last start (Citi Field) and his debacle at Yankee Stadium (more cut = less horizontal movement away from the 0 line).
Finally, vertical movement. Not as dramatic as the horizontal, but, again, some type of change.
Before we jump happily into confirmation mode, as that middle chart looks to support Warthen’s remarks, some comps are in order.
Biting into the Big Apple
Ken Takahashi and Brian Stokes both pitched in relief of Santana on Sunday, and we happen to have another game (May 11) from Citi Field where all three pitched. If you look back at the charts above, you can find the May 11 game.
Using May 11 and June 14 only, here are the speed and movement components for all three pitchers. Click to enlarge.
Horizontal spin movement
Vertical spin movement
As you can see, all three pitchers have a shift in the all-important horizontal movement. For the lefties, Takahashi and Santana, it adds “cut,” but, for the right-handed Stokes, it adds “tail.” Also, the difference is not the same for each pitcher.
Takahashi 0.5 inches Stokes 2.0 inches Santana 2.8 inches
They also lost some “rise” on their fastballs. Santana’s fastball may have been a touch slower, while the relievers may have dialed it up a notch.
This is getting complicated already, and we haven’t even looked at release points or acceleration, which also will get involved in the “adjustment” process. Theoretically. In any case, it does look like Santana, who has the highest pitch total of the three, had the most change in fastball movement between the two games. I’m not convinced that’s bankable, but, there may be indications that Santana’s fastball added around one inch of cut (or lost an inch of tail, depending on your perspective).
As I said, the attempt to handle the differences would be halfhearted. There’s much more work to be done.
Back to home
Warthen’s statement doesn’t necessarily imply the problem just cropped up in the Bronx. Stripping out road games, to avoid the mess of park effects, may help clear things up. Granted, it is not safe to assume Citi Field hasn’t had any tweaking, so another round of comps is in order. And that’s where I’ll leave it for now.
Some post-event data isn’t the only benefit, but a working larger sample of pitchers, games and pitches will help matters. And I’ll also start digging into some of the other components mentioned, but not covered, today.
References & Resources
PITCHf/x data from MLBAM.