3 rating lists are calculated using BayesElo program:

  • Engines in each season are considered same despite potential updates.
  • Each version is rated separately
  • All versions in all seasons treated as same

Simulations run 100k tournaments using ratings from seasonal list. % shown is how often engine ended in that position. All tiebreak rules are observed, however no additional crashes are assumed.

Note that ratings and simulations often have big error bars at start of division when there are new engines not seen earlier during this season and same version did not play in previous season(s).

Old posts are at https://tcecbayeselo.blogspot.com/