It's disengenious in the extreme
Do you seriously expect people to believe that you don't think creating shots for others isn't a valuable skill?
I'm sure you get this because it's pretty obvious and you're just trying to make some kind of argument but obviously the ball ..and data model... don't care how the shot is created.. because sparkling play and obvious 5 yard pass a) both still create a shot and b) even out over time across the model .. ie if one player could only do b in easy situations then they'd be out produced by a player that could do a and b
As Knight has said you either accept reversion to the mean on a proven data model or you deny it / don't accept the model
You can't be into it when it suits your argument and not into it when it doesn't
Based on this season in the league only (which people keep doing for some reason) yes you'd expect his assists to drop a bit next season per 90 and his goals to go up a bit per 90 next season .. but who really cares - they're all elite numbers however you slice it
As for fluffing Darwin (?!) ... this is going round and round.. the problem with trying to unpick his production for people that don't rate him is he's really really good / productive at everything you'd want a forward to do .. whether it's top line (actual goals and assists) or underlying shots/goals/shot creation/assists .. or just goals.. or whatever the fuck you like
I mean maybe if you did one for a scatter plot for scruffy vs stylish he'd do badly but other than tha.t...
Shot creating actions is not the same metric as assists, I was asking about assists and their value in assessing a players productivity.
I'm not coming to you from a position of eye test overrides analytics, it might be convenient for you to try and boil it down to that but its not what im trying to do. What I am trying to do is dig in to the analytics and assess their actual usefullness and application. As they're often presented as facts and inarguable, which i sometimes dont believe to be the case. If it irks you that I question how these metrics are applied, or undermine them then I'm sorry but your issue is with the metrics not being robust enough to stand up to scrutiny rather than with me for scrutinising them.
I dont care if he's scruffy or stylish, makes no odds to me.
But if someone is presenting goals + assists as a metric and calling them elite numbers, I'm within my rights to ask 'ok but how?' am I not?
Assists as a metric, is sub optimal for assessing a players creativity is it not? And optimal is what analytics is going for, cutting out noise, cutting out subjectivity and looking beneath the surface. Assist as a metric does none of that, it's a highly context dependant stat when it comes to saying "this player is adding value in attacking situations" because a final pass in a sequence which leads to a goal is not necessarily the most valuable action in that sequence. Right?
So when I see people provide G+A per 90 as a means of assessing Darwin (or any forward's) output it seems like 'bad' analytics, lazy analysis, metrics which are often applied when they align with a predetermined position that 'I think this player is good, so I'll overlook how shaky these metrics are and employ them regardless, because they support my view".....which kind of runs contrary to the entire point of analytics in football isn't it?
Bad data, or rather sub optimal or incomplete data, is just as bad as no data isnt it?
Just saying, there appears to be a lot of usage of sub optimal data presented as if it isnt.