by L-Jimmy Fri Jul 02, 2021 5:44 pm
Dttragic2016 wrote:
Oh wow, didn't realise there was so much interest in here about this stuff. Just to clarify, Sport Tech Daily is a company I launched last year and I oversee the stats/modelling stuff. Site is sporttechdaily.com & the data stuff is on edge.sporttechdaily.com/nrl
A couple of the articles explaining our work & how to use Edge are below too. As I mentioned, happy to chat further in the Draftstars thread & PMs whenever, always feel guilty posting too much on the main thread haha
https://sporttechdaily.com/analytics/predictive-analytics-applied-to-rugby-league-looking-at-try-scorers-in-the-nrl/
https://sporttechdaily.com/analytics/how-to-get-the-best-out-of-edge-in-2021/
Interesting, thanks.
A couple of questions leap to mind.
1 - this is a player-only model. Seems like opposition, team mates, team form etc could be useful?
2 - I'd be hella tempted to horserace a lot of these variables, and composite multi-variable models, to help determine variable-player weightings.
3 - There's lot of weight placed on exponential smoothing. This is tricky with messy variables like tries, and poor for trends like minutes played. Tempting to look for bayesian models.
Fun stuff though! Thanks for posting, and I'll keep following!