As you basically say, a lot of that noise should be fairly even across the league - things like injury checks, brush backs, arguing calls, seventh inning stretch, home run reviews, commercials, etc. It would be a LOT of work to strip that out.
But I don't things like mound visits should be stripped out - that is a key component of why some games take longer than others. If Posada is coming out to talk to the pitcher twice an inning, that slows down the game and is something that should be factored into the analysis.
I do think we can see, on average over the course of the season, whether a team's offense or pitching is more "at fault" for game times by looking at some time stamp data. But I think a lot of the noise would have to be assumed to be fairly similar across the league, otherwise you could spend a LOT of time trying to work the data.
Let's take that 10 inning game the other night. There were 6 pitching changes which add around 2m30s (I pulling this out of my ass) each (likely more with the manager's slow walk out and the actual switch time). That's 15 minutes there. Say there were an additional 15 visits in that game between the pitching coach, infielders and catchers at 45s average, which adds 11m15s. Between the two that's 26 minutes of added time to the game that you're factoring into the time it takes pitchers to make pitches.
Just pulling that out drops that time from 39s/pitch ((201x60)/308) to 34s/pitch (((201-26)x60)/308).
That's a pretty significant change. If you were to remove the time between each half inning, think of what that does to the suggested time per pitch. That's true of every other factor that is outside of the true time it takes: Pitcher receives the ball from catcher, does his routine, throws pitch (or pickoff, or pitchout). I can see where catcher visits to discuss strategy/signs makes sense to be added in, but IMO that is typically the catcher deciding to visit the pitcher and isn't really the fault of the pitcher.
I did some rough number crunching and came up with OWHIP (Offense's Walks + Hits per Inning Played) for the Sox and the Padres to get a high win team and a high lost team to compare. These could be slightly off, but the Sox had a OWHIP of 1.53. The Padres had a OWHIP of 1.34. The Yankees' OWHIP was 1.58.
I don't have the data on average time per game for these teams, but I'm pretty sure it correlates to the data if you were to add the two teams OWHIPs together and compare it to the average time per game of those matchups.
Or it's just pointless data that let me waste some time at work today that only says something about a team's OBP that we already know.
Edited by EastCoasterOutWest, 09 April 2010 - 02:39 PM.