Given the rumors about Ozan Kabak to Liverpool, I thought I'd do a brief thread on his isolated impact, which has always made me curious, given how highly rated he is.

Ozan Kabak is currently estimated as very bad at both creating xG and suppressing opposition xG.

Kabak's first season in my database is 18/19 when he moved to Stuttgart from Galatasaray. After the 18/19 season, my model estimated him as being "bad," but actually slightly better than his current estimates.

Being just "bad" at 18 in a new league is not unconscionable.

The main confounder for these estimates, though, is that Schalke and Stuttgart, though to a slightly lesser extent, have been extremely bad at driving play during his times with them.

For example, this season, every single player pairing at Schalke has horrid xG differentials.

Given that the model can only estimate a player's impact based off of their team's xG differential when the player is on the pitch v. not on the pitch, playing on a team that never achieves a + xG differential in any circumstance will hurt a player's estimated impact.

I do not consider that aspect in isolation to be a weakness of the model; bad teams largely have bad players. But, if all of the players that Kabak plays with are horrible, there is the possibility that their weakness is getting stapled to Kabak due to the way regressions work.

So, lets look at some of the players Kabak has played a sizeable number of minutes with.

All of them are also estimated as bad.

Stambouli, though, who Kabak shares 80% of his BL minutes with, is at least somewhat significantly better.

There is the possibility that Kabak's teammates are so bad that their negative impact is being partially attributed to Kabak. For example, Oczipka's play driving could be so bad that some of the impact is misattributed to Kabak. Conversely, it could actually be Kabak's fault.

If Kabak was "good," there is enough variation in Schalke's personnel choices (Kabak has played with 28 different outfield teammates in 20/21) this season that there would be at least some sort of an uptick in xG differential that would be attributed to Kabak. But, there is not.

With that being said, though, given how bad his common teammates are, Kabak may not actually be as bad as my model estimates. He is almost certainly bad, at the moment, but he could just be a below-average defender, rather than a horrible one.

I'll end the rambling with this: It is certainly possible he improves in the future. Playing on a terrible team could be preventing him from playing to his true ability or strengths. Playing for Liverpool would be a totally different prospect for a defender than Schalke.