Does your model for the Alleycats use their outdoor stadium in your projected +/-? I feel like I comment every week, but this is all really interesting and well-thought out. If we are adding to the weather data wishlist, I would hypothesize that player availabilities are also more favorable for home games. My mind wanders to 538's raptor modeling for basketball that calculates player contributions in expected wins, albeit I would hope availability is better for the playoff race. Unsure if your previous posts about roster consistency could split conditionally to Home/Away and what a correlation like that looks like or how it would affect your modelling, but it is certainly fun to wonder without doing any of the heavy lifting haha. Thanks again for the post!!
I wanted to include team-specific stadium effects in the model, but it turns out they are marginal at best since the indicator for home vs away takes care of the real signal. So, no, AlleyCats using their outdoor stadium this week isn't taken into account because it's assumed that it affects the game equally for both teams. At least the models I ran didn't seem to care much about which home stadium it was.
Love your ideas about RAPTOR and using roster consistency for it in the future. I've been thinking throughout the season about doing a RAPTOR/LEBRON (BBall Index) or plus-minus based models for individual contribution like there are in the NBA to be able to build up a more robust team strength value (and not just using RPI or some win percentage based metric), but like you said, there's a lot of heavy lifting in there that I'll try to take on in the offseason in time for the 2024 season. I'm honestly slightly shocked how long it takes to get these very basic analyses and subsequent blog posts done every week haha. Hopefully things are a good step more thorough next season!
Also, thanks for the feedback, ideas, and kind words! Feel free to add whenever. Helps me get the juices flowing.
Does your model for the Alleycats use their outdoor stadium in your projected +/-? I feel like I comment every week, but this is all really interesting and well-thought out. If we are adding to the weather data wishlist, I would hypothesize that player availabilities are also more favorable for home games. My mind wanders to 538's raptor modeling for basketball that calculates player contributions in expected wins, albeit I would hope availability is better for the playoff race. Unsure if your previous posts about roster consistency could split conditionally to Home/Away and what a correlation like that looks like or how it would affect your modelling, but it is certainly fun to wonder without doing any of the heavy lifting haha. Thanks again for the post!!
I wanted to include team-specific stadium effects in the model, but it turns out they are marginal at best since the indicator for home vs away takes care of the real signal. So, no, AlleyCats using their outdoor stadium this week isn't taken into account because it's assumed that it affects the game equally for both teams. At least the models I ran didn't seem to care much about which home stadium it was.
Love your ideas about RAPTOR and using roster consistency for it in the future. I've been thinking throughout the season about doing a RAPTOR/LEBRON (BBall Index) or plus-minus based models for individual contribution like there are in the NBA to be able to build up a more robust team strength value (and not just using RPI or some win percentage based metric), but like you said, there's a lot of heavy lifting in there that I'll try to take on in the offseason in time for the 2024 season. I'm honestly slightly shocked how long it takes to get these very basic analyses and subsequent blog posts done every week haha. Hopefully things are a good step more thorough next season!
Also, thanks for the feedback, ideas, and kind words! Feel free to add whenever. Helps me get the juices flowing.