Hugo Lloris had his detractors last season. It didn’t start well with a highly publicised drink-driving incident. Then the French goalkeeper picked up a thigh injury that kept him out for the whole of September, a run where Tottenham suffered losses to Watford, Liverpool and Inter Milan.
During the season there were several high profile blunders. For example, the rush of blood that led to him seeing red against PSV, a mistake that almost certainly cost us the game and at the time looked to have seriously dented our Champions League hopes. Then there was the error that gave Barcelona a head start in the first two minutes of our match at Wembley. Add to that the gift he gave to Jurgen Klopp’s title chasing Liverpool when we deserved at least a point at Anfield.
But the problem with assessing goalkeepers is that you remember the mistakes much more than the saves.
On Target Expected Goals
One more objective way of measuring a goalkeeper’s performances is to look at how many goals they conceded, versus the number of on target shots they faced. However, this is fairy crude – not all shots are equally hard to keep out.
A better approach is to look at “On Target Expected Goals”. This concept is similar to the regular expected goals statistic but only shots on target are considered. Furthermore, the part of the goal that the shot was heading for is taken into account – shots into the corner are usually harder to save than those aimed at the centre of the goal.
(Note: See the work of Colin Trainor on Statsbomb, where I first read about this idea).
If we divide the expected goals value of shots faced, by the actual goals conceded, we get some idea of how well a keeper performed. A larger score indicates better performance.
Here are the top 10 keepers in this season’s Premier League based on my model (note: I only include keepers who faced a minimum of 8 xG):
Well that is a bit of a surprise. Lloris comes out top!
The list seems to pass a sense check – Alisson, who is second, was lauded for his efforts in Liverpool’s title challenge. David De Gea, who was a little off form by his high standards, turns up in 10th.
Of course there are many drawbacks to this approach, not least that the data available for keepers is generally pretty crude. Then there is the question – are big errors in high profile games more important than generally solid performances? Still, it does look like rumours of Lloris’ demise are a little premature.