Trying to rate and compare goalkeepers with data is a double-edged sword. First and foremost goalkeeping data is one of the few collections of data available in football that strips away a lot of the noise normally occurring on the pitch. A shot is taken, and the crazy super computers are able to ignore everything else on the pitch, and purely examine the position quality of the shooter as well as the quality of the shot’s placement. Its relatively cut and dry, which is so hard to find in football, and why this can be a valuable exercise.
The problem then lies, to me, in the aggregation of this data. Its one thing to look at a shot and see the data confirm the eye test that a keeper made an impressive save (or the opposite – as I don’t think the data is perfect yet), but aggregating that over an entire season or half a season leaves room to draw potentially hasty conclusions because of sample size.
First, let’s take a look at how Serie A’s keepers have fared this season:
Not looking good for Musso so far! The data is suggesting that he has stopped 2.6 less goals than a keeper would be expected to based on the quality of shots that he faced. Marco Sportiello hasn’t fared a whole lot better this year, but there is an interesting name up at the top of this list, potentially someone waiting in the wings!
Not to get sidetracked, but purely looking at this data it is easy to just assume that Musso has been poor this year, and even one of the worst shot stoppers in the league. While an argument COULD be made like that, I think it is important to dig a bit further into all the individual shots he’s faced to draw a real conclusion. Otherwise we’d be drawing a conclusion based on only 42 shots that he had a chance to influence. To me, that doesn’t constitute a large enough sample to draw meaningful conclusions.
First let’s take a look at the 12 goals that Musso has conceded this season, sorted by difficulty of stopping:
Of these 12 goals, there’s a few that one could call Musso out on, but it would have more to do with his anticipation or concentration rather than the acrobatics he’s able to pull off in goal. Valeri’s goal was Musso’s fault for other reasons, and he may have been able to do better against Dia’s attempt. Otherwise, the Danilo and Vlasic goals that are rated so poorly by post-shot xG feel like vastly undervalued attempts.
I just rewatched it; the Vlasic shot especially was taken at point blank range and fired with such velocity off the cross bar, that Andre the Giant still may not have been able to save it even if he was positioned in the perfect horizontal location. That’s a little side bar on the metrics, and how they are like not perfect yet, but generally speaking the shots that Musso conceded average out to 50/50 chances. Nothing stands out as truly egregious, even though Musso is at fault for some of the areas he lacks in and around his area.
Where the data really starts to get loopy is the shots that he has saved this season. Spoiler alert, most of them have been pretty dang easy to save. Take a look:
It shouldn’t be a surprise that saved shots tend to have a lower chance of being converted than actual goals, but something tells me that there is too wide a variance between the shots Musso has saved versus the ones he hasn’t. Over half the chances he saved had less than a 10% chance of being converted, and there is no goodness in the opposite direction for him to make up ground to offset the easy chances he has had to save.
A quick comparison to Marco Sportiello, and on average the shots that Atalanta’s back up had to make were about 23% more difficult to stop.
At the end of the day measuring goalkeeper shot stopping effectiveness is a counting game, and keepers sometimes need to get lucky in their own right to help pad the stats. Sportiello got to feast in the sweet spot range of shots that have a 20%-40% chance of being converted. These are challenging shots, but ones that a keeper can save enough to quickly raise the PSxG tally in his favor. In over double the minutes, Musso only had 6% of his shots saved fall in this range (although he did permit three goals in this range; as did Sportiello), and he just hasn’t had the opportunities to make up ground on his adversaries in Serie A.
The next step would be doing a league wide comparison, I just don’t have the webscraping chops to pull shot data for each game. However as an additional proxy to see if Musso’s 6% of save opportunities is low, I can use Atalanta’s shots against opposition as a quick comparison. 36% of Atalanta’s shots that were saved (18 of 50) fell in this range. That may be another point in Musso’s favor for poor luck so far, but I’d love to do a league wide study to fully confirm that.
Hence the challenge of trying to rate his performance with only 42 data points. Has he been good? Not really. But an equally valid question could be; has he had the opportunity to be good? Against you could say not really. Right now it just feels like Musso is at the blackjack table in an unfortunate position of running off far too many 15’s and 16’s against dealer 18’s, 19’s, and 20’s. Averages will tell us he should regress back to his true talent level, but we’ll just have to wait and see.
Personally if Atalanta’s defense continues to force defenses into weak shots that never give Musso the opportunity to make up ground in this weird game of addition, fine by me. I still feel that Musso is an above average shot stopper, and his true talent is what he displayed at Udinese and during the European competitions last year. Let’s just hope he’s ready to step up and put a worldy of a save in when we need him – because I imagine we’ll need at least one make or break save from him this season. Until that time, as always, Forza Dea!
