Tactics Journal

by Kyle Boas

Analyzing football tactics

The goalkeeper should be replaced if we don't score

July 11, 2023 — There is a strange phenomenon where people overlook or cast too much of a spotlight on a goalkeeper based on the number of goals their team scores, not concedes.

The perfect example is André Onana versus Kepa Arrizabalaga. Here is a basic comparison.

Inter Milan scored 71 goals last season, and Chelsea scored 38 goals.

Figure 1.1 - Goalkeeping per 90 comparison via FBref.
Figure 1.2 - Advanced Goalkeeping per 90 comparison via FBref.

Kepa was the better shot-stopper statistically last season, with a PSxG+/- of 0.18 per 90. They both had nearly identical save percentages, yet Kepa faced 0.75 more shots per 90. Onana has the upper hand in the cross-stopping department though, stopping 0.5% more crosses attempted.

Figure 1.3 - Passing per 90 comparison via FBref.

If Onana played for Chelsea last season, would you be happy if your club paid €55m for him? I don’t think a club would pay that amount for Kepa right now, but if you put him on Inter Milan, then maybe they would.

Should Onana start for Inter Milan next season, or should they rush to replace him? That’s the question many Chelsea fans are asking about Kepa; they want him replaced.

The scapegoat for poor attacking returns is the defense, and whether they’re performing well or not, the first person to be blamed is the goalkeeper. Both made errors, but André Onana and Kepa Arrizabalaga performed like world-class goalkeepers last season.

Why did Thomas Tuchel, Graham Potter, Bruno Saltor, Frank Lampard, and now Mauricio Pochettino all say ‘Kepa is still Chelsea’s number one goalkeeper’? Because he’s been performing well. There was and is no reason to replace someone who’s performing at that high a level.

If you want to go to an extreme, look at Ederson for Manchester City. He had a career-worst 62.3% save percentage from only 2.20 shots per 90, with a -4.8 PSxG+/-. But they scored 94 goals, so no one notices.

I reckon Manchester City would have wrapped up the Premier League three weeks sooner if they had a better-performing goalkeeper who is a comparable passer, allowing them to rest their players for the final matches of the season. Alternatives exist.

A defense made up of Ruben Dias, John Stones, Nathan Aké, Manuel Akanji, and Kyle Walker should set records, not concede 39 goals.

I’m not saying Ederson is a bad goalkeeper; he’s in the S tier. However, he has been going through a rough patch recently. He started performing better towards the end of the season and was a completely different goalkeeper in the Champions League compared to the Premier League.

Goalkeepers are built on form. You need to have a short-term memory when evaluating them, or you’ll get stuck behind the curve. All three will dip and peak at different times.

I’d like to see what the reaction will be if Andre Onana performs the same as he did at Inter, but United score close to or fewer goals than last season (58). Will they shine the spotlight brighter on him, or will they be able to ignore the errors?

It’s illogical to judge a goalkeeper’s performance based on the offensive output of a team. They shouldn’t be replaced if the team can’t score goals.

Back to top Share on Twitter Email this post Copy link