TCEC - Thanks Martin and congrats to the winner Komodo

[Modern Times]

So Martin, how about a special TCEC chess960 match, Houdini 4 vs Stockfish ?

Generate random opening positions using Kirill's tool:
http://kirill-kryukov.com/chess/tools/opening-sampler/

use that as an openings database in ChessGUI, and job done !

[Dr.Wael Deeb]

May be it's time to wonder what makes TCEC an awesome event :

- Nice layout (board, menus, game score, ranking, ...)- Easy way to follow the games/tournaments, html page
- Live chat
- Latest (even beta) versions of engines
- Top and average engines
- Qualification/elimination is fair
- Wide and short book, engines have to show their strength in opening
- Long time control allow deep thinking and so, interesting games
- Engines updated/debugged between stages
- Poll to know spectators opinions
... more ?

Thanks a lot Martin !

I am not so confident of your evaluation of the TCEC event regarding point 1 & 6....

Personally,I like the classic FIDE wood layout for a chess borad or the metal layout but that's just me....

Regarding the opening lines that were used,the case was not completely flawless as we all know....

I hope to se a perfect season 3 next year....
Dr.D

[Modern Times]

Regarding the opening lines that were used,the case was not completely flawless as we all know....

It will never be flawless unless it is chess960

[Dr.Wael Deeb]

Regarding the opening lines that were used,the case was not completely flawless as we all know....

It will never be flawless unless it is chess960

A perfect truth I'd say....

[M ANSARI]

Actually there are a few Chess 960 positions that are know to be lost or at least with huge advantage from the start, but if the colors are reversed then both engines should have the same chances. I do think that this would be testing something quite different than normal chess, but it would be interesting nevertheless. Generally speaking the top Chess 960 engine is usually also one of the top Chess engines. Personally I think Komodo would be a great Chess 960 engine if it would get the extra code to accept the start positions.

[Modern Times]

Actually there are a few Chess 960 positions that are know to be lost or at least with huge advantage from the start, but if the colors are reversed then both engines should have the same chances.

I'm not convinced of that myself. The starting positions are identical for white and black, so all you have is the white advantage, and I don't believe that can ever be immediately "lost" for black or a huge disadvantage.

[Martin Thoresen]

I am not so confident of your evaluation of the TCEC event regarding point 1 & 6....

Personally,I like the classic FIDE wood layout for a chess borad or the metal layout but that's just me....

Regarding the opening lines that were used,the case was not completely flawless as we all know....

I hope to se a perfect season 3 next year....
Dr.D

For point 1, I think he is talking about the website as a whole, not the design of the chess board.

As for point 2, there will never be "perfect" openings because everyone is always complaining, in one way or another.

[Dr.Wael Deeb]

I am not so confident of your evaluation of the TCEC event regarding point 1 & 6....

Personally,I like the classic FIDE wood layout for a chess borad or the metal layout but that's just me....

Regarding the opening lines that were used,the case was not completely flawless as we all know....

I hope to se a perfect season 3 next year....
Dr.D

For point 1, I think he is talking about the website as a whole, not the design of the chess board.

As for point 2, there will never be "perfect" openings because everyone is always complaining, in one way or another.

There is no perfect openings but there are reasonble ones

The design of the websight is ok,but the colours of the chess board are not my cup of tea....

All in all a great chess initiative that could be improved even more....
Cheers,
Dr.D

[Vinvin]

May be it's time to wonder what makes TCEC an awesome event :

- Nice layout (board, menus, game score, ranking, ...)- Easy way to follow the games/tournaments, html page
- Live chat
- Latest (even beta) versions of engines
- Top and average engines
- Qualification/elimination is fair
- Wide and short book, engines have to show their strength in opening
- Long time control allow deep thinking and so, interesting games
- Engines updated/debugged between stages
- Poll to know spectators opinions
... more ?

Thanks a lot Martin !

I am not so confident of your evaluation of the TCEC event regarding point 1 & 6....

Personally,I like the classic FIDE wood layout for a chess borad or the metal layout but that's just me....

Regarding the opening lines that were used,the case was not completely flawless as we all know....

I hope to se a perfect season 3 next year....
Dr.D

Book is a matter of taste.
For the layout, I'd say it's the best one currently to broadcast computers tournament. Compare it to http://www.clemens-keck.de/livegames/ , the tournament is even more interesting (24 CPUs, 3 top engines) but it's less funny to follow (old chessboard, no PV, no auto refresh, no graph, ...) .

[Haikouichthys]

I'm working on a solution to the openings issue that factors in 1) play balance, 2) apparent draw avoidance, 3) commonness and 4) difficulty, defined as the degree to which the leading engines disagree, adjusting their outputs in accord with their respective evaluation scaling much as you propose.

Expressing these four factors as percentile ranks corresponding to each opening position, it will then be an easy matter to weight each category differently as we proceed through the five stages of the competition and use the consolidated figure to select openings. Difficulty, for example, would be unimportant in Stage 1 but be quite important in the Superfinal. Commonness would be desirable in Stage 1 but become progressively less important. Play balance and draw avoidance would be relatively constant in importance through the stages, though draw avoidance is less important in Stage 1 and more important in the Superfinal.

At the moment I'm assembling the data. It'll be a close-run thing to have everything in place prior to the start of next season but I feel confident the grand design will be in place. It's quite an undertaking; the foundation is over 44,000 GM-produced positions after eight moves.

Much more detail to follow when the project is completed.

This sounds great, but... eight moves? Has Martin told you something that he hasn't told the rest of us or have I missed something that Martin has said?