Sorry man. Lots of good work. But you're not accounting for the team getting to the top of the order the 2nd or 3rd time against the starter. That also explains the end-game spike in whiff rates - the closer coming in.
I include both times through the order and SP/RP as fixed effects. I'm also directly controlling for quality of stuff, via aStuff+. I'm also limiting the sample to innings 1-8, so there are no "closers" (although that's not the reason I cut the 9th and extras).
I have further discussion on working out those specific biases in my first post. But I'll add that, because the position of the sun moves so much in six months, the distribution of the sun is not as big of a concern as you might think.
This does a fantastic job of explaining/justifying light as the main factor of underperformance of batters at T-Mobile. Do you think that if the batter's eye were changed it alone would be enough to significantly reduce the upper hand pitchers have at the stadium?
Also, this is inarguably somewhat unrelated, but do you have any advice for modeling the distance a baseball travels given it's launch angle/exit velocity? i'm struggling to figure out how to code such a thing and im not sure i can just ignore things like wind resistance.
My favored culprit is the batter's eye, whether that be the color or angle (or both). It could certainly be something else about the stadium, like the lights, but given that the batter's eye is so prominent, I figure it must be involved somehow. I'm not sure what the fix would be, or if one is even possible. But I do think changing the batter's eye could change the effect and possibly make the stadium more neutral.
As for modeling batted ball distance, I'm not sure! I've considered starting down that path, but I'm also not a physicist, and I'd be a bit out of my element there. If you're looking for an existing calculator, Alan Nathan has an excellent one. I believe the conclusions from his work is what's featured on Baseball Savant, as well.
Sorry man. Lots of good work. But you're not accounting for the team getting to the top of the order the 2nd or 3rd time against the starter. That also explains the end-game spike in whiff rates - the closer coming in.
I include both times through the order and SP/RP as fixed effects. I'm also directly controlling for quality of stuff, via aStuff+. I'm also limiting the sample to innings 1-8, so there are no "closers" (although that's not the reason I cut the 9th and extras).
I have further discussion on working out those specific biases in my first post. But I'll add that, because the position of the sun moves so much in six months, the distribution of the sun is not as big of a concern as you might think.
This does a fantastic job of explaining/justifying light as the main factor of underperformance of batters at T-Mobile. Do you think that if the batter's eye were changed it alone would be enough to significantly reduce the upper hand pitchers have at the stadium?
Also, this is inarguably somewhat unrelated, but do you have any advice for modeling the distance a baseball travels given it's launch angle/exit velocity? i'm struggling to figure out how to code such a thing and im not sure i can just ignore things like wind resistance.
Thank you for your time!
My favored culprit is the batter's eye, whether that be the color or angle (or both). It could certainly be something else about the stadium, like the lights, but given that the batter's eye is so prominent, I figure it must be involved somehow. I'm not sure what the fix would be, or if one is even possible. But I do think changing the batter's eye could change the effect and possibly make the stadium more neutral.
As for modeling batted ball distance, I'm not sure! I've considered starting down that path, but I'm also not a physicist, and I'd be a bit out of my element there. If you're looking for an existing calculator, Alan Nathan has an excellent one. I believe the conclusions from his work is what's featured on Baseball Savant, as well.