Stockfish, the open source engine by Marco Costalba, Joona Kiiski, Gary Linscott and a huge community of contributors, has convincingly won TCEC Premier Division. Stockfish collected 61,0/84 points to secure the gold medal and the right to challenge the current champion Houdini for the TCEC Grand Champion title.
The Superfinal match of the Top Chess Engine Championship starts this Thursday at 17:30 CET and you can follow the games live on the official site
The reigning champion Houdini by Robert Houdart finished clear second with 54,5/84. It had better performance than the winner against Komodo, but a higher draw rate against lower rated engines.
Both Houdini and Stockfish incurred only a one loss during the Premier division. Stockfish lost a game against the bronze medalist Komodo, while Houdini lost a direct match with Stockfish. Scroll down for the full standings.
The Superfinal details
The TCEC S11 Superfinal will be 100 games match aka x50 double round robin. This match is played with 50 different openings so that each engine plays both black and white of the same position. The match will be presented with opening 1 used in games 1 and 2, then opening 2 used in games 3 and 4 etc. If the match is theoretically won for one side before game 100, the match will still continue until all 100 games have been played. In the case of a drawn match there will be a rapid match of 16 games with a time control of 25′ + 10″ with random openings selected from earlier in the same Season. In case it is still tied there will be a Blitz match of 8 games with a time control of 3′ + 2″.
The winner of the Superfinal will be crowned the TCEC Grand Champion and will keep this title until there is a winner in the next Superfinal. There is no automatic qualification for the reigning Grand Champion, it will have to go all the way through the next Season for it to be able to defend the title.
Stockfish will have an updated dev version for the Superfinal. The team has added a few extra improvements and are coming in high spirits in a quest for the TCEC Grand Champion title, something that has not been in the hands of Stockfish since Season 9.
Robert Houdart from Houdini has not been able to release a new version for the Superfinal, but still believes in the current title holder, despite recognizing the advance of Stockfish.
The match is relatively short and anything can happen in the 100 games sprint. History has shown that early winners keep their advantage to the end of the match. However, statistics are a different story, the real one will start this Thursday live on the official site
S12 to start after the Superfinal
As TCEC is now a continuous divisions system, S12 is going to start right after the Superfinal. Read more about it here
TCEC Premier Division final standings
1. Stockfish 61,0/84
2. Houdini 54,5/84
3. Komodo 51,5/84
4. Fire 42,5/84
5. Chiron 36,0/84
6. Andscacs 35,0/84
7. Fizbo 31,0/84
8. Ginkgo 24,5/84
The big three is dominating the standings again, although every single engine has a different position than last season. Fire is the clear winner of the other contenders, ahead of Chiron and Andscacs, which completed it’s quest to promote and stay in the Premier Division. Fizbo and Ginkgo get relegated, with the status of favorites in First Division.
Assaf Wool from MyTCECexperience provides detailed statistics on the Premier Division.
Draw rate, wins
Final draw rate was 61.0%
The three most common game termination causes were:
37.8% – TCEC win rule
36.3% – TCEC draw rule
14.9% – 3-fold repetition
There were no engine crashes in the division, there was one game where the runner software crashed (game 268, Ginkgo – Houdini). The game reached a drawn 6-man position and was manually set as a draw.
Moves per game
The first double RR was bookless, and a summary of the bookless openings can be found in a previous post. Here I summarize the results of the remaining games, with book sequences by Cato. The first letter of the ECO codes was distributed as follows:
If we use the opening ‘family name’ (using format FAMILY_NAME: VARIANT….) the top 3 are:
Sicillian – 58 times
English – 24 times
King’s Indian – 24 times
The opening variant was mostly determined by the book sequence. 96.4% of the game pairs repeated the same ECO code twice, and 87.1% repeated the same opening variant. In all of the game pairs the ECO first letter was the same in both games.
Reverse pairs, wins
Again I’m restricting to the games with book sequences.
There were only 2 game pairs with a 1-1 biased result.
Reverse pairs, same moves
Here I’m also restricting to the games with book sequences.
Pairs of reverse games diverged very quickly, 51.8% of the pairs diverged immediately after book, 78.6% repeated at most 2 plys.