The best model fit to womens football data seems to be this model based on the Negative Binomial.

They make a DC version of the Negative Binomial distribution. And they look at some other Poisson and Negative Binomial based models. Unfortunately, no plots of these models are shown.

I have often wondered if there is any way to generalize or extend the Dixon-Coles model, and it turns out there is! Based on some theory from the 1960's, in this intersting paper the authors study a lot of different extensions! arxiv.org/abs/2307.02139

Another feature is the possibility to convert probabilities to odds with a specific margin.

A tutorial with a thorough description of all methods ca be found here.

It works for 2, 3 and also multiple outcomes, so can be used for player vs player sports (football, hockey etc) and racing.

Multiple "winners" is also possible, for example odds for reaching final.

implied: Convert Between Bookmaker Odds and Probabilities Convert between bookmaker odds and probabilities. Eight different algorithms are available, including basic normalization, Shin's method (Hyun Song Shin, (1992) &lt;<a href="https://doi.org/10.2307%2F...

My #rstats package for converting bookmaker odds to proper probabilites, by removing the bookmaker margin, can be found here.

8 methods are available, taken from academic literature and from articles online.

calibration curves showing uncalibrated and calibrated match win probabilities of women's soccer matches. data source: FiveThirtyEight

✍️new blog post on using the tidymodels probably package to calibrate model predictions. #rstats

tonyelhabr.rbind.io/posts/probab...