Let me give you a preview of something I'm working on, and call for suggestions!
This is a line from an Excel spreadsheet.
There were 2097 pitches thrown in last Sunday's AL games. I'm not wild about GameDay's pitch recognition, which seems to vary from park to park. I think the main problem comes with fastballs - sometimes they get broken down into fastball variations (sinkers and cutters) - sometimes they're all just fastballs. Sometimes changeups and fastballs get confused. I was at this afternoon's game, and Felix Hernandez throws his changeup harder than Shaun Marcum throws his fastball, so you can see how that would happen from time to time.
I do find, by the way, that the GameDay velocity readings are invariably the same or within 1 mph of those shown on the Rogers Centre screen in left field (and that means they can be several mph slower than what you may see on your television, depending on who's doing the broadcast.)
As a happy bonus, Sunday's AL slate had some interesting things to offer. We'll be able to look at some strange pitch selection courtesy of Kevin Slowey and Joe Mauer, some very impressive work by Erik Bedard and Roy Halladay, and some very high velocity work by Joba Chamberlain and Craig Hansen. We have a game winning walk-off grand slam in one game, a Mariano Rivera sighting in another... and best of all, one of the rarer beasts in the game: a pitcher striking out the side on nine pitches.
Anyway - 2097 pitches, all entered into a spreadsheet. Pitch Type and Velocity information is available for 2077 of the 2097 - not perfect, but not bad. And I've got several days to play with the information.
For example: there were 537 at bats in those games, and therefore there were 537 0-0 counts. In 368 of those at bats (68.5%), the pitcher threw a first-pitch fastball. If a batter is going to swing at the first pitch, he's much more likely to go after a first pitch fastball. The batter went after the first pitch 113 times, took a first strike 109 times, and watched ball one go by 146 times.
But he was not going to go after a first pitch curveball. There were 39 of them, and only 14 missed for ball one. But the hitter took strike one 23 times, and fouled off the other two. None were put in play, none were swung at and missed.
I can quite happily do this sort of thing all day, but I thought I'd put this up first to see if there were any other ideas out there.
This is a line from an Excel spreadsheet.
3a 40 40 13 3 Halladay TOR R Huff BAL L 2-2 3 1 13 0-2 Curve 77 s KI went through the GameDay Logs for all seven of last Sunday's games in the AL and made an entry like the above for every pitch. The first five numbers are internal house-keeping stuff, so I can keep everything in place. If you're curious, the first number is the game (Game 3, away team batting), the second is the number of pitches thrown by the team in the game (40), then the number thrown by the pitcher in the game (40), in the inning (13), and in the at bat (3). The Halladay and Huff columns don't really need explanation. After that we have the pitch situation: the score (2-2), the inning (3rd), the number of outs (1), the men on base (runners on 1st and 3rd), and the count on the hitter (0-2). Then we have the pitch that was thrown, its velocity, and what the hitter did. I'm using my STATS pitch codes here, "s" means swing-and-a-miss, and I trust you all know what K means.
There were 2097 pitches thrown in last Sunday's AL games. I'm not wild about GameDay's pitch recognition, which seems to vary from park to park. I think the main problem comes with fastballs - sometimes they get broken down into fastball variations (sinkers and cutters) - sometimes they're all just fastballs. Sometimes changeups and fastballs get confused. I was at this afternoon's game, and Felix Hernandez throws his changeup harder than Shaun Marcum throws his fastball, so you can see how that would happen from time to time.
I do find, by the way, that the GameDay velocity readings are invariably the same or within 1 mph of those shown on the Rogers Centre screen in left field (and that means they can be several mph slower than what you may see on your television, depending on who's doing the broadcast.)
As a happy bonus, Sunday's AL slate had some interesting things to offer. We'll be able to look at some strange pitch selection courtesy of Kevin Slowey and Joe Mauer, some very impressive work by Erik Bedard and Roy Halladay, and some very high velocity work by Joba Chamberlain and Craig Hansen. We have a game winning walk-off grand slam in one game, a Mariano Rivera sighting in another... and best of all, one of the rarer beasts in the game: a pitcher striking out the side on nine pitches.
Anyway - 2097 pitches, all entered into a spreadsheet. Pitch Type and Velocity information is available for 2077 of the 2097 - not perfect, but not bad. And I've got several days to play with the information.
For example: there were 537 at bats in those games, and therefore there were 537 0-0 counts. In 368 of those at bats (68.5%), the pitcher threw a first-pitch fastball. If a batter is going to swing at the first pitch, he's much more likely to go after a first pitch fastball. The batter went after the first pitch 113 times, took a first strike 109 times, and watched ball one go by 146 times.
But he was not going to go after a first pitch curveball. There were 39 of them, and only 14 missed for ball one. But the hitter took strike one 23 times, and fouled off the other two. None were put in play, none were swung at and missed.
I can quite happily do this sort of thing all day, but I thought I'd put this up first to see if there were any other ideas out there.
