If you’re behind on the La Liga prediction model we’ve been using this season, there’s really no better place to catch up on it than our recent podcast with Charlie Tuley (@analyticsLaLiga). You can find that podcast here. Follow Charlie and Michel (@msocanalytics) on Twitter for all the great things they are doing with football numbers. Michel’s page includes this same prediction model for several other leagues.
In matchday 3, the prediction model was outstanding, getting 7/10 match predictions correct.
A great showing from the model this week - went 7/10 on results and only 1/10 on scorelines. Hopefully we can ride this high through next matchday!! pic.twitter.com/pVP9yZduTz— Charlie Tuley (@analyticsLaLiga) August 29, 2021
This brings the overall total of correct result predictions to 14/30, which, when you consider the fact that the model began the season using data entirely from last season (before the transfer market) I would say their results are outstanding. Yes, their total is still less than 50%, but we are in a sport where three different results are possible from a match, and oftentimes the second most likely result in their prediction has been the one that actually happened. Some of those matches are like Villarreal’s last match with Atleti, where the model predicted Atleti’s control of the game even though the score ended up tied.
Villarreal, Sevilla, Barcelona, and Alaves are all sitting this one out because of a combination of scheduling issues from the international break and upcoming UCL fixtures, so we have just 8 matches this week. After this matchday, 80% of the data being used in the algorithm will be from this season, and I expect with that to see an uptick in average accuracy.
Celta and Elche are the biggest underdogs this week, will they be able to upset the apple cart? I look forward to seeing how this latest round of predictions stacks up.