Karpov-Kasparov WCC (1984) Game 9 position

Karpov-Kasparov WCC (1984) Game 9 position

Postby guyhaw » Sat Apr 15, 2017 2:06 pm

I don't have the PC resources to evaluate this with FINALGEN - any volunteers?

The position is about to feature in a new book about B-v-N. The theory is that the only possible defense is 66. ... Bh1 but it is not clear if even this is good enough.

8/4N3/8/1p6/1PkP4/P3K3/6b1/8 b - - 0 66

Thanks in advance,

Posts: 6
Joined: Sat Apr 15, 2017 2:02 pm

Re: Karpov-Kasparov WCC (1984) Game 9 position

Postby guyhaw » Sat Apr 29, 2017 8:05 am

In fact, sorted now: 8/4N3/8/1p6/1PkP4/P3K3/6b1/8 b - - 0 66 is a draw after only-move 66. ... Bh1.

After the actual 66. ... Bb7??, FinalGen says 'dtf' = 35m. A line I followed hits 81. Nxa4+ after which dtm = 25m.

Would be interested in hearing what positions, with 8 men or more, Finalgen users have analysed (or would like analysed).

Guy Haworth
Posts: 6
Joined: Sat Apr 15, 2017 2:02 pm

Re: Karpov-Kasparov WCC (1984) Game 9 position

Postby ChemChess » Fri Jun 16, 2017 8:20 pm

Also getting zero for an evaluation...FinalGen is amazing.

To illustrate how powerful it is, I compared the results from the deep analysis of the position using the latest build of Stockfish, which has the highest ELO of any program on the planet right now (aside from the assembly-language based Stockfish clones, of course).

Using an overclocked i7-4790k @ 4.7 GHz in the Fritz 15 GUI, I analyzed the position to depth 104 (50+ plies deeper than other computers/clusters on the ChessBase cloud). It was confirmed that the position converges to an evaluation of approximately 0! (See pic below). Keep in mind that this evaluation also utilized the FULL 6-man syzygy table-bases as a reference. I also manually probed a few promising seven-piece positions that Stockfish simplified the position down to using the subscription-based Lomonosov 7-man tablebases. Again, the evaluation was confirmed in each case to be 0.


This is incredible considering I can get the SAME result from FinalGen with half the time and a fraction of the processing power it took to brute-force this endgame using some of the latest software/hardware available.

FinalGen has incredible potential. I'll hopefully report back with more findings soon.
Posts: 3
Joined: Fri Jun 16, 2017 7:55 pm

Re: Karpov-Kasparov WCC (1984) Game 9 position

Postby jkominek » Sun Jul 16, 2017 7:17 am

After some Youtube wanderings I found that the Canadian-Russian chess author Roman Jiganchine evaluated this position using FinalGen a couple years ago, writing about the results in an April 5, 2015 entry on his blog site http://roman-chess.blogspot.com.

Searching for the label Karpov brings the entry to the top of the search results list.
http://roman-chess.blogspot.com/search/ ... -%20Karpov

The FinalGen screenshot reports 1h:7m:45s of clock time and 126.8GB of required working space to find its answer (on whatever computer Jiganchine was using at the time).

By any chance is it John Nunn who is working on an endgame book of 8-man positions?

Posts: 1
Joined: Sun Jul 16, 2017 7:00 am

Re: Karpov-Kasparov WCC (1984) Game 9 position

Postby guyhaw » Tue Oct 30, 2018 10:22 am

JK ... thanks for the reference - I was not aware of this previous work but it's good to see it being done, and confirming my computation.

No, the Karpov-Kasparov (1984) position was sent to me by Karsten Mueller and i think features in his forthcoming book 'Understanding Minor Piece Endgames'.

Posts: 6
Joined: Sat Apr 15, 2017 2:02 pm

Return to FinalGen EndGame Generator Forum

Who is online

Users browsing this forum: No registered users and 1 guest