With a CL spot secured, the BuLi season is done for us, and Harry Kane has racked up 35 goals so far. Obviouslee (Kane's favorite word), strikers do not operate in a vacuum; they need service and support from the midfield. But which of our midfielders was most conducive to enabling Kane goals this season? That's the question I was curious about. Everyone will have their own analysis, and hence opinion, and it will be difficult to agree on this. To quantify contributions, one could simply look at the number of assists, but this would ignore a large number of intangibles. There are too many to name. What I propose instead is a simple, more comprehensive metric. For each midfielder, count the number of minutes they shared the pitch with Kane, tally up the Kane goals scored in these periods, and normalize by minutes. Let's call this expected Kane goals, or xKG for short. The outcome of this exercise is quite surprising, at least to me.

First, there are two standout players, Laimer and Kimmich, who occupy similar roles in midfield but differ substantially in terms of xKG. Who would have thought that Laimer is our player with the highest xKG by far? This is truly unexpected. Secondly, there is no clear relation between midfield position and xKG. DMs, CAMs and L/RWs can be found all over the place. There is one each in the top three and also the bottom four of the xKG chart. Old-timers such as Choupo and Muller have been quite effective; one of them should perhaps be on the pitch at all times. Gnabry has had the least minutes shared with Kane (369) and could therefore be a statistical outlier. Youngsters such as Tel and Musiala have been slightly less influential when it comes to xKG. Another standout player is Pavlovi?, whose xKG is remarkable given his more defensive role. Finally, Sane shared the most minutes with Kane (2090) and also the most Kane goals (27) but his xKG is in the mid-range. Overall, there was no correlation between either minutes and xKG (r(10) = -1.66, p = 0.12) or between Kane goals and xKG (r(10) = -0.18, p = 0.86).

How should we interpret these numbers? The data is purely correlational and any causal explanation based on it will have to make many additional assumptions. I leave that entirely up to you, let me know in the comments.

Data source: transfermarkt.com, plotting and stats: Julia. Positional changes such as JK's move to RB, and changes brought about by substitutions (e.g., LG moving to RB yesterday), were taken into account. Only midfield minutes were included for each of these players.