Do you have any graphs handy that kind of replicates the one used first in the blog post but a bit less ambiguous, maybe without model grouping? I feel like it would have been a bit more fair to include proper names, and individualize them rather than group everything together by something, and then present your own model on its own.