The ratings don't diverge around 2015 and it's not due to a hype split.
They diverge around 2000 (slowly, with increasing divergence) and it's due to the fact that this is when BGG started collecting ratings.
As author points out, newer games have higher ratings. If we hypothesize that brand new games get overrated when first released, we can posit that this "newness" effect would apply for games that are released during BGG lifetime, but not before.
Using a Wilson score instead of the actual rating would control for this effect.
They diverge around 2000 (slowly, with increasing divergence) and it's due to the fact that this is when BGG started collecting ratings.
As author points out, newer games have higher ratings. If we hypothesize that brand new games get overrated when first released, we can posit that this "newness" effect would apply for games that are released during BGG lifetime, but not before.
Using a Wilson score instead of the actual rating would control for this effect.