Like most models, Estimated Contribution is built with multi-season Regularized Adjusted Plus-Minus (RAPM). But it& #39;s hard to use RAPM that spans over a decade for the WNBA. So instead I adapt the approach used by @DSMok1 and @James_Brocato for BPM 2.0: https://www.basketball-reference.com/about/bpm2.html ">https://www.basketball-reference.com/about/bpm...
Specifically, I use three sets of 5-year RAPM from 2003-2017. It& #39;s informed by a simple prior based on team quality and player workload. I also apply a "luck adjustment" to account for random variation in 3-point and free throw shooting.
The WNBA Estimated Contribution model also draws inspiration from @NateSilver538 / @FiveThirtyEight’s RAPTOR, particularly the way it constructs the on/off component.

#OnOff_Components">https://positiveresidual.com/analysis/estimated-contributions-in-the-wnba/ #OnOff_Components

I">https://positiveresidual.com/analysis/... find that the RAPTOR approach works better than standard on/off ratings.
You can follow @presidual.
Tip: mention @twtextapp on a Twitter thread with the keyword “unroll” to get a link to it.

Latest Threads Unrolled: