How many engines will Sacrifice the Rook to win?

[Chessqueen]

An (UCI-)engine canNOT resign.

Indeed, but some user interfaces do resign for the engine. (It's at least an option in both Fritz and Shredder's GUI.)

And that is very smart, since what is the point to continue after 120.....Rxd4 121. Kxd4 f5! Please check move 120.....Rxd4
https://tcec-chess.com/#div=q43t&game=294&season=21

Except... Guess what ? ClassicAra played it because it was using Syzygy Tablebases and 121. Kxd4 is a 6-man TB win. It wasn't even considering recapturing the Rook as best, since it was ending the game instantly because of TB adjudication. Case closed.

I knew the answer way before I posted this here on Talkchess, it was discussed in TCEC, and I gave several hints, I even post the link here, but finally you were the only one who caught it,https://tcec-chess.com/#div=q43t&game=294&season=21 ===> http://talkchess.com/forum3/posting.php ... 2&p=892484

[Alayan]

Check move 120.......Rxd4! clever way to force the opponent engine to resign ==>
https://tcec-chess.com/#div=q43t&game=294&season=21

The only reason this was played is because this led to an immediate 6-men TB win and the engine hadn't found the mate. These sort of simplification into TB wins are very common in engine chess, but the moves involved are not brilliant as in they don't make the win any easier or more certain. It also has nothing to do with being "materialistic".

A good benchmark for "equally good moves" (preserving win/draw with perfect play) is how they fare with less-than-perfect play. By this benchmark, Rxd4 is a poor move because the position rook up is extremely easy to play. Without the rook, white's king can't take the back passed pawn without the more advanced one running to promotion, but it's not any easier, and a beginner would be more likely to mess it up.

[mvanthoor]

I knew the answer way before I posted this here on Talkchess, it was discussed in TCEC, and I gave several hints, I even post the link here, but finally you were the only one who caught it,https://tcec-chess.com/#div=q43t&game=294&season=21 ===> http://talkchess.com/forum3/posting.php ... 2&p=892484

Even so, Rxd4 is not the best move in this position. It is the move where winning the game can be guaranteed by the table-base, because it's in a position where it knows everything. After Rxd4, the table-base has perfect information.

You asked if we knew an engine that would sacrifice the rook on d4. We don't know one, because none would. Then in the end you 'reveal' that you already knew that the resulting position comes from the 6-man tablebase, which means that it isn't actually the engine that chose to play that move; any engine that has access to the 6-man table-base _should_ play Rxd4, because in that position, the engine just gets told, by the table-base: "You are winning, and I will tell you EVERY MOVE you will need to play in EVERY position from now on..."

This has completely nothing to do with engines that are 'less materialistic".

Humans actually do the same thing. If I can trade down to a King + Bishop + Knight vs. King endgame, or I can trade down to King + Rook vs. King endgame, then... guess what.

The K+B+N endgame is technically "better", because you have 1 point of material extra compared to K+R, but I am not confident that I can actually checkmate with K+B+N vs. K... I *AM* confident I can checkmate with K+R vs. K, so I will trade down to the "less good" endgame. The engine is doing exactly the same, not because it's less materialistic, but because the table-base says: "If you play Rxd4, I can tell you EXACTLY how to win. (And with any other move, I can't.)"

edit: Alayan basically says the same thing.

So if you knew the entire game beforehand, including the fact that the engine had access to a 6-man table-base, you're trolling because you asked a completely non-relevant question ON PURPOSE to waste everyone's time. I think I'm going to put you into my ignore list.

[mar]

Even so, Rxd4 is not the best move in this position.

not only is it not the best move, it's actually the worst move that leads to the longest mate

[Chessqueen]

Even so, Rxd4 is not the best move in this position.

not only is it not the best move, it's actually the worst move that leads to the longest mate

Actually by giving the Rook away most other engines that use 6 men EGTB will eventually resign quicker, and you will not even see the shorter mate, but instead a quicker ending game, whereas, if the best move was played the other engine will paly for at least another 4 moves before resigning. A shorter mate doe NOT necessarily mean that you will ever see it being executed between engines

NOTE: Go ahead use the same position after move 119 and feed it to stockfish vs Komodo both using 6 men EGTB, and I will guarantee you that after playing the best move, the game will be prolonged for at least another 4 moves before resigning. And once again the main post was not asking which move lead to a shorter mate, please go back to the original post where I mentioned that specifically. There is no need to continue talking about this, you all know what I posted originally and continue looking for the shortest mate or the best move

[Guenther]

Even so, Rxd4 is not the best move in this position.

not only is it not the best move, it's actually the worst move that leads to the longest mate

Actually by giving the Rook away most other engines that use 6 men EGTB will eventually resign quicker, and you will not even see the shortest shorter
mate, but instead a quicker ending game, whereas, if the best move was played the other engine will paly for at least another 4 moves before resigning. A shorter mate doe NOT necessarily mean that you will ever see it being executed between engines :roll:

NOTE: Go ahead use the same position after move 139 and feed it to stockfish vs Komodo both using 6 men EGTB, and I will guarantee you that after playing the best move, the game will be prolonged for at least another 4 moves before resigning. And once again the main post was not asking which move lead to a shorter mate, please go back to the original post where I mentioned that specifically. :roll:

You still have a grand misconception about 'resigning'... The GUI adjudicates (it is not even resigning), even better said the admin/tournament director let the GUI adjudicate on behalf of the engines, according to 'his gusto settings'.
It has nothing to do with the engines or engine names you like to drop like pets here. If the TD once sets adjudication off nothing happens before draw by rules or mate, but this would of course show how useless this thread was from the beginning.
Mentioning the 'main post' is hilarious too, because your so called main posts are moving targets, because you never know all the shortcomings
in your original starter posts, thus they need to be doctored all along (or let's just say say it: blown up to a full troll).

Sure we never know, if you intentionally lie, or pretend to lie to hide your stupidity as long a thread evolves, but there are several indications you would like to be more known for being a liar, instead of an unteachable noob in this forum. Thus it is very likely the second part is true for you.
(Being that dumb about computerchess after reading here for 20 years is an enormous acchievement though.)

[Rebel]

Even so, Rxd4 is not the best move in this position.

not only is it not the best move, it's actually the worst move that leads to the longest mate

[IQ_QI]

Hello Chessqueen and others,
I am sorry that I am the cause of this problem.
CrazyAra was only recently ported to classic chess, and I did not expect ClassicAra to be invited to TCEC Season 21.
I have tested the engine so far mainly in crazyhouse and with game adjudication enabled based on centi-pawn evaluation.
In these scenarios, the behaviour seen is not displayed.
I plan to let ClassicAra play more meaningful moves in such positions with Syzgy tablebases enabled in a future version.

Kind regards,
Johannes

[Chessqueen]

Even so, Rxd4 is not the best move in this position.

not only is it not the best move, it's actually the worst move that leads to the longest mate

Actually by giving the Rook away most other engines that use 6 men EGTB will eventually resign quicker, and you will not even see the shortest shorter
mate, but instead a quicker ending game, whereas, if the best move was played the other engine will paly for at least another 4 moves before resigning. A shorter mate doe NOT necessarily mean that you will ever see it being executed between engines

NOTE: Go ahead use the same position after move 139 and feed it to stockfish vs Komodo both using 6 men EGTB, and I will guarantee you that after playing the best move, the game will be prolonged for at least another 4 moves before resigning. And once again the main post was not asking which move lead to a shorter mate, please go back to the original post where I mentioned that specifically.

You still have a grand misconception about 'resigning'... The GUI adjudicates (it is not even resigning), even better said the admin/tournament director let the GUI adjudicate on behalf of the engines, according to 'his gusto settings'.
It has nothing to do with the engines or engine names you like to drop like pets here. If the TD once sets adjudication off nothing happens before draw by rules or mate, but this would of course show how useless this thread was from the beginning.
Mentioning the 'main post' is hilarious too, because your so called main posts are moving targets, because you never know all the shortcomings
in your original starter posts, thus they need to be doctored all along (or let's just say say it: blown up to a full troll).

Sure we never know, if you intentionally lie, or pretend to lie to hide your stupidity as long a thread evolves, but there are several indications you would like to be more known for being a liar, instead of an unteachable noob in this forum. Thus it is very likely the second part is true for you.
(Being that dumb about computerchess after reading here for 20 years is an enormous acchievement though.)

Herr Günther Warum bist du so unhöflich? Du zeigst deine Klasse wirklich und redest wie eine Person aus einem Ghetto in diesem Forum.
Mr. Guenther, and I do NOT know if you deserve to be call Mr. Obviously like always you are the one that Are an untouchable noob and always so rude, here is the reason just posted by the Author of CrazyAra

"CrazyAra was only recently ported to classic chess, and I did not expect ClassicAra to be invited to TCEC Season 21.
I have tested the engine so far mainly in crazyhouse and with game adjudication enabled based on centi-pawn evaluation.
In these scenarios, the behaviour seen is not displayed.
I plan to let ClassicAra play more meaningful moves in such positions with Syzgy tablebases enabled in a future version."

Kind regards,
Johannes
Top