https://theanalyst.com/articles/tottenham-thomas-frank-stats-premier-league-expected-points
Tottenham’s underlying numbers are as bad as their Premier League position suggests. There is little reason to believe Thomas Frank can turn this around and save Spurs’ season.
Small section .. full article at link above
Spurs have won the ‘xG battle’ (generated more expected goals than their opponents) in just eight of their 22 Premier League games this season. Only Burnley, Fulham and West Ham have won on xG fewer times than them.
What’s more, of Tottenham’s eight xG victories, four have come at home when they have gone behind and been chasing the game, and their opponents have dropped back to defend their lead – against Liverpool, Fulham, Aston Villa and Manchester United. Spurs lost the first three matches in that list and drew with United.
These numbers do highlight one limitation of xG. Spurs might have won on xG against Liverpool and Fulham, for example, but on both occasions that was largely the result of them going 2-0 down and fighting their way back into the game late on against a team who were happy to sit on their lead. Expected goals doesn’t take into account game state.
Opta’s expected points model runs into similar problems, but it is useful nonetheless in trying to work out where teams ‘deserve’ to be in the league. The model simulates the number of goals scored by each side in a match based on the xG value of every shot taken, and then simulates every match 10,000 times and assigns points to either team from the proportion of those sims they won, drew and lost. The idea is that over a period of time, xG data can tell us how well or badly a team is creating chances, thus giving a better indication than the actual table of how everyone is really playing.
Throughout Frank’s reign, Spurs have persistently overachieved compared to their underlying numbers. After nine games of the season, for example, Spurs were third in the Premier League, but the xPts table had them down in 13th. In other words, their performances were worthy of enough points to be 13th in the league. No team had a bigger difference between their actual and ‘expected’ positions. This graphic shows that, rather than the 17 points they actually had, their underlying numbers suggested they deserved 11.2 xPts.
Knowing what we know now, these were warning signs from back in October that things were going to head south. And so it proved, as Spurs now find themselves 14th in the Premier League (while they are still 13th in the xPts table). They have sunk to a position their underlying numbers suggested was on the cards months ago. The table right now is a far better reflection of Spurs’ displays.
Tottenham’s Champions League Position Cannot Save Doomed Thomas Frank
Tottenham’s underlying numbers are as bad as their Premier League position suggests. There is little reason to believe Thomas Frank can turn this around and save Spurs’ season.
Small section .. full article at link above
Spurs have won the ‘xG battle’ (generated more expected goals than their opponents) in just eight of their 22 Premier League games this season. Only Burnley, Fulham and West Ham have won on xG fewer times than them.
What’s more, of Tottenham’s eight xG victories, four have come at home when they have gone behind and been chasing the game, and their opponents have dropped back to defend their lead – against Liverpool, Fulham, Aston Villa and Manchester United. Spurs lost the first three matches in that list and drew with United.
These numbers do highlight one limitation of xG. Spurs might have won on xG against Liverpool and Fulham, for example, but on both occasions that was largely the result of them going 2-0 down and fighting their way back into the game late on against a team who were happy to sit on their lead. Expected goals doesn’t take into account game state.
Opta’s expected points model runs into similar problems, but it is useful nonetheless in trying to work out where teams ‘deserve’ to be in the league. The model simulates the number of goals scored by each side in a match based on the xG value of every shot taken, and then simulates every match 10,000 times and assigns points to either team from the proportion of those sims they won, drew and lost. The idea is that over a period of time, xG data can tell us how well or badly a team is creating chances, thus giving a better indication than the actual table of how everyone is really playing.
Throughout Frank’s reign, Spurs have persistently overachieved compared to their underlying numbers. After nine games of the season, for example, Spurs were third in the Premier League, but the xPts table had them down in 13th. In other words, their performances were worthy of enough points to be 13th in the league. No team had a bigger difference between their actual and ‘expected’ positions. This graphic shows that, rather than the 17 points they actually had, their underlying numbers suggested they deserved 11.2 xPts.
Knowing what we know now, these were warning signs from back in October that things were going to head south. And so it proved, as Spurs now find themselves 14th in the Premier League (while they are still 13th in the xPts table). They have sunk to a position their underlying numbers suggested was on the cards months ago. The table right now is a far better reflection of Spurs’ displays.
