Premier League 21 Jan 2022

Chelsea v Tottenham

After some midweek drama in one or two of the rescheduled Premier League games, the weekend arrives with more football (it’s also possible that all the postponements are the reason our model below doesn’t quite show all the matches).

Chelsea having failed to beat Brighton, now have a tough match against Tottenham, who may be on a high after their last minute goals against Leicester in their midweek game. One match can make all the difference between whether a manager is suddenly seen as ‘genius’ (Conte) or somehow ‘failing’ (Tuchel).

For this game though, there are some consistencies. Our model gives home advantage to Chelsea, it also takes note that Chelsea’s expected goals are much higher than Tottenham’s (almost ‘1’ whole goal more). Chelsea have an average of 2 goals per game, whereas Tottenham have only 26 from 19 matches.

In terms of predictions, our model predicts a 1-0 home win. In terms of likely winners our model matches up quite closely to Bookmakers Mean Odds (see table below for Outcome Probs and Mean Odds), with less than 3% separating the two models for a likely Chelsea win (59.4% to 56.8%).

Everton v Aston Villa

This is the early kick off on Saturday. The reason I’ve picked this game though is not so much that, or that Duncan Ferguson will be managing Everton, it’s more that there is a difference in predicted outcomes.

Villa have spent a fair bit in the transfer window, acquiring Digne from Everton and Coutinho on loan (who scored last weekend on his debut), Everton have also bought new full-backs but with a new manager it’s interesting that our model favours Everton at 43.4%, when none of this is taken into account, only past form, ELO ratings, home advantage and recent results.

Bookmakers mean odds slightly favour Villa, but only by 1.3% which really leads to the conclusion our model predicts which is a likely 1-1 draw. This exact score is 12.2% likely.