Tuesday, January 15, 2008
Pitch F/X and Joba Chamberlain
In 2007, MLB decided to enhance their Gameday application. As a numbers junkie (aka statdork), the best thing that's come out of that in my opinion is the rolling out of detailed pitch by pitch data. Data for 2007 is incomplete as it's been rolled out slowly across MLB, but it now appears to be available in most of the stadiums and should be everywhere for 2008.Pitch F/X is the name of the system that MLBAM is using to track the detailed pitch information. They record things like starting velocity, ending velocity, release point, break, and all other kinds of stuff. Having the data available is great, but figuring out how to get it and make use of it looked like a royal pain in the ass.
Thankfully, someone else did all the work for me as far as figuring that out. Over at the blog Fast Balls, Mike Fast detailed how to build a pitch by pitch database. I used Fast's instructions and have been downloading and loading the data for the past week or so. This is great stuff and I thank Fast for making his database building instructions open source.
There's a ton of data in here that I don't fully understand yet, and as I mentioned before it's pretty incomplete because the system has taken a while to roll out across MLB, but I think some data is better than none.
I am hoping to make use of this data in some of my posts in 2008, especially when it comes to Joba Chamberlain, Phil Hughes, and Ian Kennedy. Today I'd like to look at Joba specifically.
We all saw how Joba burst onto the scene with his high 90s fastball and nasty slider, throwing 24 innings of 1192 ERA+. He does have a curve and changeup, but neither was used very much as a reliever. For Joba, Pitch F/X had detailed pitch information for 225 of his 334 pitches.
The first and easiest thing to look at would be the breakdown of his pitches thrown.
Fastballs: 139
Sliders: 79
Changeups: 7
Next, here are the high, low and average velocities for those three pitches.
Fastball
High: 101
Min: 92.9
Avg: 97.3
Slider
High: 90.4
Min: 82.5
Avg: 86.6
Changeup/Curve (I can't tell which is which from the data)
High: 84.5
Min: 79.4
Avg: 77.2
Lastly, here's a look at what the end result has been for the pitches we have data for.
| Pitch Type | Result | # |
| Fastball | Ball | 45 |
| Ball In Dirt | 1 | |
| Called Strike | 30 | |
| Foul | 26 | |
| Foul (Runner Going) | 0 | |
| In play, no out | 6 | |
| In play, out(s) | 19 | |
| In play, run(s) | 2 | |
| Swinging Strike | 10 | |
| Swinging Strike (Blocked) | 0 | |
| Pitch Type | Result | # |
| Slider | Ball | 27 |
| Ball In Dirt | 2 | |
| Called Strike | 9 | |
| Foul | 7 | |
| Foul (Runner Going) | 1 | |
| In play, no out | 0 | |
| In play, out(s) | 5 | |
| In play, run(s) | 0 | |
| Swinging Strike | 24 | |
| Swinging Strike (Blocked) | 4 | |
| Pitch Type | Result | # |
| Changeup/Curve | Ball | 4 |
| Ball In Dirt | 0 | |
| Called Strike | 2 | |
| Foul | 0 | |
| Foul (Runner Going) | 0 | |
| In play, no out | 0 | |
| In play, out(s) | 1 | |
| In play, run(s) | 0 | |
| Swinging Strike | 0 | |
| Swinging Strike (Blocked) | 0 |
I'm not sure how accurate this data is yet, or how consistent it is from park to park. I'd imagine that this year's data will be fine-tuned a bit based on last year. I love the potential for new information that this technology will give us though.
Thanks again to Mike Fast for his open-source instructions on parsing this data, it was a huge help.
Update: Thanks to mmiller for pointing out this link to Josh Kalk's work. Thanks to Kalk for his work and data as well.
Page 1 of 1 pages: