Now how about finishing. I measure it as the ratio of goals to expected goals. If a player is a good finisher, we would expect them to score more goals than their expected goals would suggest. Here there is some persistence but it is much weaker.

We might worry that a forwards' expected goals depend on their context (teammates, tactics etc...) as well as their own ability. The chart below shows that expected goals are persistent even for forwards who move teams, the relationship is slightly weaker but still there.

First, let's look at expected goals, which measures the quality of shooting chances a forward takes. Forwards with a high expected goals in one period of matches are likely to continue to have high expected goals in the next period of matches (even when looking at short periods).