Very good read. However, the chess example about the Tata Steel tournament is incorrect.

The prediction chart shown was not based on guesses from grandmasters. It was simply the implied probability of each player winning based on their relative Elo ratings, their current tournament standing, and upcoming pairings.

Essentially it was a Monte Carlo simulation of the tournament, which was updated after each round by the author on reddit.

Doesn’t impact the validity and strength of this article, but just wanted to point that out.