Thanks, super interesting. Does seem like “should reasonably be expected” is doing a lot of work, especially when you consider there is an understandable amount of outcome bias with how a big chance is given.
Excluding penalties, the conversion rate is about 38% in the Premier League (which immediately calls into question the idea that the player 'should' score!).
Salah was at 35.1% last season in the league. If we then go backwards through the seasons, he was 42.3%, 40%, 48.4%, 29.6%, 46.2%, 42.3% and 46.3%. He's never been super hot but also only one season (2020/21) markedly below par.
Mane had 57.7% (2018/19) and 51.9% (2021/22) seasons, while Díaz was 54.2% last season, for instance.
Good stuff Andrew. Because data alone can miss important context, the clips were particularly valuable to judge the shot quality for myself. Appreciate the work you put into this. The data on finishing matches matches the "eye test" in recent seasons, i.e. he's an average finisher. Beyond finishing, there's also been a big decrease in the # of "quality" np shots and that's not all on Mo.
An elite forward needs to take lots of shots from the danger zone (6-yd box extended to 18 yards). I like looking at shot count from the danger zone bc forwards of Mo's ability can often score shots that an "average finisher" can't. It gives you another data point along side np xG. In the EPL this season, Mo's getting 1.4 shots per 90 inside the danger zone (vs. 2.0 & 2.4 the previous two seasons). That's 30% fewer shots from a good location vs. last season. Yikes!
He's also down in np xG shots worth 35% or more per 90. A 35% probability is my unscientific definition of a "good chance". Mo's only had 0.10 such chances p90 this season (vs. 0.56 and 0.60 the previous two seasons). That's an 80% decrease!
Some of this decline is down to Mo but a lot is the shortcomings of other factors, e.g. the tactics and the team around him.
I keep an eye on danger zone shots but for the team rather than individual players - I assume you've compiled that manually? If there's an available source then let me know!
And as the average conversion rate for a big chance is around 37%, your unscientific approach looks pretty spot on!
As with anything, there is more than one explanation. But the supposed best chances Salah has had this season have frequently done him few favours.
Thanks, super interesting. Does seem like “should reasonably be expected” is doing a lot of work, especially when you consider there is an understandable amount of outcome bias with how a big chance is given.
I’m curious what the league/European conversation rate of ‘Big Chances’ is. And where Salah ranks in the last few seasons.
Excluding penalties, the conversion rate is about 38% in the Premier League (which immediately calls into question the idea that the player 'should' score!).
Salah was at 35.1% last season in the league. If we then go backwards through the seasons, he was 42.3%, 40%, 48.4%, 29.6%, 46.2%, 42.3% and 46.3%. He's never been super hot but also only one season (2020/21) markedly below par.
Mane had 57.7% (2018/19) and 51.9% (2021/22) seasons, while Díaz was 54.2% last season, for instance.
Good stuff Andrew. Because data alone can miss important context, the clips were particularly valuable to judge the shot quality for myself. Appreciate the work you put into this. The data on finishing matches matches the "eye test" in recent seasons, i.e. he's an average finisher. Beyond finishing, there's also been a big decrease in the # of "quality" np shots and that's not all on Mo.
An elite forward needs to take lots of shots from the danger zone (6-yd box extended to 18 yards). I like looking at shot count from the danger zone bc forwards of Mo's ability can often score shots that an "average finisher" can't. It gives you another data point along side np xG. In the EPL this season, Mo's getting 1.4 shots per 90 inside the danger zone (vs. 2.0 & 2.4 the previous two seasons). That's 30% fewer shots from a good location vs. last season. Yikes!
He's also down in np xG shots worth 35% or more per 90. A 35% probability is my unscientific definition of a "good chance". Mo's only had 0.10 such chances p90 this season (vs. 0.56 and 0.60 the previous two seasons). That's an 80% decrease!
Some of this decline is down to Mo but a lot is the shortcomings of other factors, e.g. the tactics and the team around him.
Really great comment, thanks.
I keep an eye on danger zone shots but for the team rather than individual players - I assume you've compiled that manually? If there's an available source then let me know!
And as the average conversion rate for a big chance is around 37%, your unscientific approach looks pretty spot on!
Yep, manual tracking of danger zone shots ... wish there was a source!
Definitely, I'm surprised there isn't really given how important shots in that region are.
Danger zone counts looked great last night 8-1 and our shots were clustered around the 6 with 4 of very high quality. Best performance this year