It’s Easier for Alisson: Evaluating the Liverpool Goalkeeper’s Champions League Final Saves With AI Simulation

By: Kevin Chroust and Paul Power | June 5, 2019

You’ve likely already read that Alisson made more saves with a clean sheet in a Champions League final than anyone since Porto’s Vitor Baía helped Jose Mourinho to his first Champions League title in 2003/04. And you’ve likely seen Jürgen Klopp discuss how Alisson makes stopping shots look so easy. But there’s little out there actually quantifying his value as a shot-stopper because there’s little out there that can meaningfully measure it.

Alisson made eight saves Saturday with his night slowly going from unbothered to busy as time ran out on Tottenham, but how reliant on Alisson were Liverpool for those saves compared to an average goalkeeper?

“He makes the hard things look easy,” Klopp said. “I’m not sure he gets enough credit.”

It’s difficult to always give keepers credit when so much of it is based on our opinion, so here we’ll take a look at the saves Alisson made by using historical data to compare him to average keepers.

It was a 2-0 final, but if we consider expected goals it was much closer – 1.43-1.30 in Liverpool’s favour. But even expected metrics don’t quite capture what we’re after here with keeper performance. Few will argue traditional goalkeeper measurements such as clean sheets and save percentages are truly indicative of performance. Expected saves only measure how a player performed against the shots they faced. Keepers may have completely different types of saves to make depending on the defensive style of their team and the opponents they face. So rather than using metrics that might not capture these variables, we’ve developed a method to simulate each goalkeeper for shots beyond those they faced themselves, then compare who would concede the fewest goals. To do this, we must be able to accurately simulate how one keeper would manage against another keeper’s shots – or better yet, how each keeper would perform against every shot in a uniform sample.

With this new model, we can show the likelihood of a goal for each shot if it had occurred against an average Premier League keeper, and we can also show the likelihood of a goal against Alisson. Essentially, we’re showing how much assurance Alisson provides compared to normal goalkeeping.

We’ve spent the past two months publishing weekly keeper rankings across various leagues, and Alisson ended the season as the top Premier League keeper when simulated against a uniform sample of shots rather than just his own. ESPN wrote about this method in advance of the Champions League final.

For the eight shots Alisson saved in the final, the expected goal value for a standard goalkeeper was 0.76. For Alisson, it was a quarter goal below that (0.51). Four of the shots he faced would have had an expected goal value of 0.10 or higher (10 percent likelihood of scoring) for the average keeper, but only two did for Alisson.

Those two shots came one after the other as Heung-Min Son struck a low, hard shot from distance to Alisson’s right, followed by Lucas Moura’s weaker shot from just above the penalty spot. For those two chances, Alisson’s expected goals against happened to be the same as the average keeper, meaning those aren’t areas where he differentiates himself as a shot-stopper. Son’s came with a 0.10 xG, while Moura’s was 5 percent higher (0.15).

He handles each very well, parrying the first wide of an approaching attacker, then collecting the second without a rebound, but as we stated, these aren’t chances on which his particular skill set above the average keeper was displayed.

Where, then, did Alisson’s separation from other keepers come though? Christian Eriksen’s direct free kick in the 85th minute. For the average keeper, this shot had a 0.14 xG. For Alisson, it was hardly a bother at 0.03. He parried it well off target and out of play without a rebound, whereas against another keeper, it was nearly five times as likely to result in a goal. The 0.11 xG difference was the highest on any shot he faced.

We can also simulate how things might have played out differently had Alisson been in the Tottenham goal for a key moment—for example, when Liverpool sealed the match with their second goal.

Below, we’ll see that the one area of strength Lloris has over Alisson is low and to his immediate right. For the second goal, Lloris had to go far to his left for a very well-placed shot by Divock Origi.

Against Lloris, that Origi shot came with a 0.35 xG. If you’ve seen that goal, you may have noticed Lloris wasn’t planted when the shot occurred but had left his feet, which makes it more difficult for him to react to the shot. With Alisson in goal, it dips to nearly half of that at 0.20.

Klopp used his subjective, expert eye to state that Alisson makes it look easy compared to other keepers. With AI-differentiated data, STATS can confirm what Klopp sees: For Alisson, it is in fact measurably easier.