Hi friends, hope you’ve been well. When trying to dive deep into the WNBA, I found a void where I wanted advanced analytics tools to be. I built a site that tries to communicate the previously-unquantified impact in WNBA basketball. This is Help the Helper https://t.co/ASfk0g1mV5
@TheStalwart We'll see this replicated to an extent with proprietary fine tuned models. Consultancies and banks have very large private corpuses of info on how to do parts of the job (ie. slides)...
@mikegrib8 Yeah, it’s a good point. You learn this naturally over time with the feedback of future higher level performance, though, which is part of that taste development
One thing I'd recommend to folks getting into prospect scouting is to watch film where your context is limited--both where you have limited priors and where you have limited access to opinions of others. This helps with developing "taste." I found HS/AAU valuable for this
Some recent Help the Helper additions: official positional designations and filters for official positional designations and # of seasons played
https://t.co/9n1icXoXZX
I think opponent shot profile is such an important defensive indicator, so I've been trying to develop some rim protection metrics. 2025 Valkyries are giving me some trouble lol
@_punkcouple_ Thank you! The base RAPM model is multi-year since single-year is crazy noisy with W possession totals. Definitely considered using a box score prior, but I wanted to avoid it for philosophical reasons and felt good about the smell test results without it...
Hi friends, hope you’ve been well. When trying to dive deep into the WNBA, I found a void where I wanted advanced analytics tools to be. I built a site that tries to communicate the previously-unquantified impact in WNBA basketball. This is Help the Helper https://t.co/ASfk0g1mV5
@ApatheticBrent Rankings are something I will continue to think about, but I am hesitant because I'm not sure I'd think about the values ordinally. Close values should be thought of as predictions where the confidence intervals have a lot of overlap. Appreciate the thoughts!
@ApatheticBrent One stat is never the end all be all, and there is plenty of noise in this one, but I think it is a solid descriptive metric of player impact within a given context...
@_punkcouple_ Thank you! The base RAPM model is multi-year since single-year is crazy noisy with W possession totals. Definitely considered using a box score prior, but I wanted to avoid it for philosophical reasons and felt good about the smell test results without it...