8-piece chess endgame tablebase – New Record!

The first results of 8-piece chess endgame tablebases are shared in a chat with developer, computer chess enthusiast, and physicist, Marc Bourzutschky. He shares analysis and findings of classic positions from Keres-Levenfish, with analysis by Huebner from ChessBase Mega Database and Kasparov-Karpov from the 1990 World Championship, as well as the new record checkmate!


  1. Sorry. 100 times 70 TB is 7,000 TB not 200 TB! What I understood wrong? Thanks.

  2. If a quantum computer could solve it quickly enough then it is irrelevant how much storage it would take. You don't have to store it. Just every time you need it, recompute it.

  3. Checkers was solved by booking up from the opening not solving the endgame. Eight pieces is still a large chunk of the game completely solved. That's why I designed 9×9 chess that is much more complicated.

  4. A couple of thoughts:

    1. Some research reveals that Kasparov-Karpov game was, in fact, adjourned after move 88 (89.Re1 was the move that was sealed).

    2. Given the enormous size of these tablebases, I wonder if Marc and others researching them have looked into tape storage. Tape storage is significantly cheaper per unit of storage, the big issue is that the tape drives themselves are quite expensive.

  5. So what Is the solution to the Keres vs Levenfish position on 22:25? Is it a draw or a win?

  6. well it is only nessesary to store the first position in 8 pieces endgames with the winning move or if it is losing, we don't need to store 500 moves combination how to win. We can always turn on the computer if we want to know how it is winning.

  7. Ken Thompson's $50 bet against Browne to win KQKR vs tablebase is legendary.

  8. The interviewer says that even if we solved chess we wouldn't have any place to put it. But that's not entirely true. I mean it's true that we wouldn't be able to store the recursive tablebases for the 32, 31, 30…. piece tables, but the 32-piece tablebase proper would just be a single entry. 🙂

  9. Well, GM Dr. Robert Hübner is really famous for his analysis without the assistance of any engines! Of course you'll find mistakes/errors in any only human analysis! 😇

    Interestingly enough Hübner's analysis can sometimes even be more accurate than Kasparov and his team (see his books on "My Great Predecessors" Volume 1-5)! 😮‍💨

Leave a Reply

Your email address will not be published.