Updated: Aug 19, 2020
While this season will be remembered for Liverpool’s perceived dominance, a closer look at the statistics suggests that they weren’t the standout performers the table suggests. Prior to the season, domestic exchanges, European bookmakers and Asian bookies all suggested that Manchester City were the most likely to come out on top after the 38 game regular season, with odds of 1.6, followed closely by Liverpool at a market-price around 3.75. Sheffield United were the bookmaker’s consensus favourites for the drop, with Norwich, Brighton, Burnley, Newcastle and Aston Villa all priced lower than 4.0 to be relegated. At first glance, the bookmakers appear to be no more than moderately accurate in their pre-season predictions. However, expected goals tells a different story entirely.
The expected goals (xG) metric has seen increased popularity in the football world over the last few years for knowledgeable fans and sports bettors building more sophisticated models than simply results-based or score-based analysis. The more sophisticated statistics measures the quality of a shot as a function of the position of the shot on the pitch, the assist type and whether or not it was a headed chance, among a vast array of other variables. Those betting on Man City to win the league will be aggrieved to know that the majority of xG calculations suggest that Liverpool outperformed their goal differential by 5 to 10 goals, while Man City under-performed their expected goal differential by one or two. If each game was settled on xG rather than actual goals scored, both Understat and FootballxG would have made Man City the winners of the premier league this season. According to xG, Chelsea and Man Utd will make up the remaining two Champions League places, which is now the favourite outcome among bookmakers.
As for relegation, most xG statistics suggest that Newcastle are the clear outliers, having outperformed their xG by a gigantic 13 to 20 points, depending on the source of this statistic. This would not only result in them being relegated but finishing bottom of the league, accompanied by Aston Villa and Norwich. While Sheffield United would have a small adjustment in their points total and finishing position, they would be safely in mid-table. Consequently, with the exception of Chris Wilder’s men, the bookmakers were exceptionally accurate in their pre-season predictions if xG was used as the deciding factor in settling games, rather than goals.