New Ratings System
Harvard data scientist and US National Chess Master Mark Glickman published research proposing a new chess rating system.
Context
Traditional chess rating systems treat draws as equivalent to half a win and half a loss. Chess ratings – which measure how strong a player is – are used to organize tournaments, award titles, and pair players fairly. These systems don’t account for the fact that stronger players draw more often. The 2018 world chess championship between Chess Grandmasters Magnus Carlsen and Fabiano Caruana saw all 12 games end in draws – the first time in history that occurred – and lead to tiebreak games. The old system led to a “rating stagnation” that didn’t distinguish between elite players or track changes in their strength.
New Approach
Last December, Glickman published research proposing the first rating system that models draw probability based on player strength.
