what is the best engine for big material handicap?

Discussion of anything and everything relating to chess playing software and machines.

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
Uri Blass
Posts: 8556
Joined: Wed Mar 08, 2006 11:37 pm
Location: Tel-Aviv Israel

what is the best engine for big material handicap?

Post by Uri Blass » Sat Oct 05, 2019 3:37 am














I would like to know result of some competition of engines against humans in these conditions when engine play the weaker side.
I would like to know if strong humans are better than the best engines in playing the weaker side(I expect them to be better but I do not know about tests that prove it).

I think to give initial rating for weak humans based on results in these conditions and the question is what is the initial rating to give for every score that is not 100%

Note that games can be played without a chess clock because humans who cannot win easily in the first position are usually not the type of humans that more time can help them and I expect a game without clocks to be finished in less than 30 minutes.

Dann Corbit
Posts: 9994
Joined: Wed Mar 08, 2006 7:57 pm
Location: Redmond, WA USA
Contact:

Re: what is the best engine for big material handicap?

Post by Dann Corbit » Sat Oct 05, 2019 4:03 am

I analyzed this one before. It is a loss in 15 moves:
acd 31; acs 7920; bm c3; ce -32737; dm -15; pm c3; pv c3 e5 g3 d5 f3 Qf6 g4 Qh4+ Kd1 Qxh2 g5 Qh1+ Kc2 Qg2 e4 dxe4 b4 exf3 a3 Be6 c4 Bf5+ Kb3 Qxd2 Ka4 Qd1+ Kb5 Bd7+ Ka5 b6+;

I think it may have been one of those Valentin positions
Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning.
But sharing ideas is an even greater virtue. We have another word for this. It is called teaching.

Milton
Posts: 121
Joined: Wed Mar 08, 2006 11:58 pm

Re: what is the best engine for big material handicap?

Post by Milton » Sat Oct 05, 2019 10:48 pm

Dann Corbit wrote:
Sat Oct 05, 2019 4:03 am
I analyzed this one before. It is a loss in 15 moves:
acd 31; acs 7920; bm c3; ce -32737; dm -15; pm c3; pv c3 e5 g3 d5 f3 Qf6 g4 Qh4+ Kd1 Qxh2 g5 Qh1+ Kc2 Qg2 e4 dxe4 b4 exf3 a3 Be6 c4 Bf5+ Kb3 Qxd2 Ka4 Qd1+ Kb5 Bd7+ Ka5 b6+;

I think it may have been one of those Valentin positions
This version of Stockfish says it's a mate in 14 moves.

New game Line
rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/4K3 w kq - 0 1

Analysis by Stockfish 160919 64 BMI2:

1.h4 e6 2.a3 Qxh4 3.d3 Qh1+ 4.Kd2 Qxg2 5.Ke1 Qg1+ 6.Kd2 Qxf2 7.b4 Nf6 8.c3 a5 9.b5 Bxa3 10.d4 Ng4 11.Kd3 Qe3+ 12.Kc2 a4 13.d5 Qxe2+ 14.Kb1 Qb2#
Black mates: -+ (-#14) Depth: 59/29 01:07:46 205331MN

lkaufman
Posts: 3678
Joined: Sun Jan 10, 2010 5:15 am
Location: Maryland USA
Contact:

Re: what is the best engine for big material handicap?

Post by lkaufman » Sat Oct 05, 2019 11:01 pm

I think that an engine would have to be specially designed to put up decent resistance in such hopeless positions. I'm pretty sure any human GM could resist better at such large handicaps than any normal chess engine. For giving rook odds (or less) the top engines (with high Contempt) are probably stronger than even Carlsen, but much beyond a rook normal chess doesn't work well anymore. I have some experience with this; I have a good friend who is about 1400 level, and he seems to be between rook odds and rook + knight odds against either me or Komodo, not much difference.
Komodo rules!

User avatar
Ovyron
Posts: 2505
Joined: Tue Jul 03, 2007 2:30 am

Re: what is the best engine for big material handicap?

Post by Ovyron » Sun Oct 06, 2019 2:57 am

I think the fact that a special purpose engine would need to be built for this proves that handicap chess isn't chess.
Great spirits have always encountered violent opposition from mediocre minds.

Uri Blass
Posts: 8556
Joined: Wed Mar 08, 2006 11:37 pm
Location: Tel-Aviv Israel

Re: what is the best engine for big material handicap?

Post by Uri Blass » Sun Oct 06, 2019 8:01 am

I disagree that it is not chess.

For me it is chess and I would like to rank humans based on their ability in it.

The fact that you need a special purpose engine show a general weakness of chess engines
that are not designed to do the best against humans in the relevant positions.

User avatar
Ovyron
Posts: 2505
Joined: Tue Jul 03, 2007 2:30 am

Re: what is the best engine for big material handicap?

Post by Ovyron » Sun Oct 06, 2019 12:51 pm

I still remember the very first involvement with chess I had in my life. I didn't even have a chess board, it was just a book for kids, with cartoons, that showed the rules of the game.

The very first thing it did was telling about the shapes and names of the pieces.

The next thing it did was showing how to put them on the board, rook knight bishop king queen bishop knight rook, with the white square of the board on the top bottom corner, and queens on squares that matched their color.

This is a core part of what makes chess, chess.

If you ignore the queen part of that, and instead have queens starting on dark squares:



That's no longer chess.

If you add an extra knight to each side:



It's no longer chess.

If you make the Knights royal (and the King a normal piece), so that "checkmating" either of the opponent's knights wins you the game:

(huh, the opening position would look the same)

It's no longer chess.

If you get rid of the queen and change it for a Camel:

Image

It's no longer chess.

If players start the games with a Knight they can drop in unocupied squares on their moves... you get the idea.

So if a player starts the game with some pieces removed from the starting board, it's clear it's also not chess. There's a name for it, "Fairy Chess."

Humans can be ranked and be very strong in chess variants, but this ranking is unrelated to chess, and chess with material handicaps is a chess variant.
Great spirits have always encountered violent opposition from mediocre minds.

User avatar
Nordlandia
Posts: 2406
Joined: Fri Sep 25, 2015 7:38 pm
Location: Sortland, Norway

Re: what is the best engine for big material handicap?

Post by Nordlandia » Sun Oct 06, 2019 1:34 pm

Ovyron: that's quite puristic view of chess as a whole.
Chess players are a strange bunch. Complaining nobody understands us, while at the same time we are very disapproving of any different way of having fun with the game.
Some players report a light feeling of nausea when they start to play chess960 and are confronted with a new start position. I'm convinced this is because they actually have to think starting from the first move, and this unfamiliar feeling literally sets their thoughts spinning. It's like the panic you feel in traditional chess when your opponent makes a move you didn't expect.
Quotes from this thread - https://www.chess.com/forum/view/chess9 ... -a-variant

lkaufman
Posts: 3678
Joined: Sun Jan 10, 2010 5:15 am
Location: Maryland USA
Contact:

Re: what is the best engine for big material handicap?

Post by lkaufman » Sun Oct 06, 2019 6:12 pm

Ovyron wrote:
Sun Oct 06, 2019 12:51 pm
I still remember the very first involvement with chess I had in my life. I didn't even have a chess board, it was just a book for kids, with cartoons, that showed the rules of the game.

The very first thing it did was telling about the shapes and names of the pieces.

The next thing it did was showing how to put them on the board, rook knight bishop king queen bishop knight rook, with the white square of the board on the top bottom corner, and queens on squares that matched their color.

This is a core part of what makes chess, chess.

If you ignore the queen part of that, and instead have queens starting on dark squares:



That's no longer chess.

If you add an extra knight to each side:



It's no longer chess.

If you make the Knights royal (and the King a normal piece), so that "checkmating" either of the opponent's knights wins you the game:

(huh, the opening position would look the same)

It's no longer chess.

If you get rid of the queen and change it for a Camel:

Image

It's no longer chess.

If players start the games with a Knight they can drop in unocupied squares on their moves... you get the idea.

So if a player starts the game with some pieces removed from the starting board, it's clear it's also not chess. There's a name for it, "Fairy Chess."

Humans can be ranked and be very strong in chess variants, but this ranking is unrelated to chess, and chess with material handicaps is a chess variant.
About thirty years ago a tournament of top GMs was held in honor of GM Polugaevsky in which all players agreed to open with the Sicilian defense (I think they even specified "open Sicilian") every game. Was this a chess tournament? Most people would say yes, even though the start position was not the normal one, because the opening was normal and no one had to participate. Other tournaments specified "king's gambit (or Evan's gambit, etc.). Were they chess tournaments? I would still say yes, although there is at least a counter-argument that those openings are rare among top players. But if we had a tournament with all games starting with White having only king and pawns, it would be hard to call that a chess tournament, even though the position is legal, since the only issue is how many moves it takes to mate, which is not a goal in normal chess. The point is that there is no hard line between "chess" and "fairy chess", it is a matter of judgment. Chess 960 isn't strictly speaking chess, since the start positions are illegal and the rules slightly different, but it still feels like normal chess. Handicap of f7 pawn is a legal position, which has occurred in many matches and tournaments (mostly in the 1800s); there are more than a hundred such games in the MegaBase database by the best players of their time, and the games feel like normal chess games after the opening. Same with "pawn and two" (f7 handicap with e4 played and WTM). Also same with knight odds (b1 or g1, usually b1), although with a piece less White generally feels compelled to play recklessly, so it starts to feel a bit different. Rook handicap seems to me the maximum that I think of as real chess, because there are very few games in the databases with larger handicaps, and because the play is just too unlike normal chess. Then there are many smaller handicaps that have few recorded games but which are otherwise pretty normal, such as a White pawn, or two moves to start, or Black opening 1...f6? and perhaps 2...Kf7? My point is that what constitutes a real chess game, tournament, or match is somewhat subjective, but should at least include serious matches played by great players of the past (meaning pawn and move, pawn and two moves, and knight odds) that are in today's databases.
Komodo rules!

User avatar
Ovyron
Posts: 2505
Joined: Tue Jul 03, 2007 2:30 am

Re: what is the best engine for big material handicap?

Post by Ovyron » Mon Oct 07, 2019 12:46 am

lkaufman wrote:
Sun Oct 06, 2019 6:12 pm
About thirty years ago a tournament of top GMs was held in honor of GM Polugaevsky in which all players agreed to open with the Sicilian defense (I think they even specified "open Sicilian") every game. Was this a chess tournament?
Let me twist it around: Suppose you run a normal chess tournament, but two GMs are friends of each other, and before they arrive, they agree that if they face each other they will play 1. e4 e5 2. Nf3 Nc6 3. Bb5 a6 4. Ba4 Nf6 5. O-O Be7 6. Re1 b5 7. Bb3 d6 8. c3 O-O 9. h3 Bb7 10. d4 Re8 11. Ng5 Rf8 12. Nf3 Re8 and get a draw by three-fold repetition. Are they still playing chess? Or would you agree they made that arrangement to avoid playing chess? To see what I mean, imagine that everyone on your tournament makes such an arrangement, and then no novelty is played, or some people play fake novelties in an arranged game that ends in a draw.

Can we agree that if that happened, instead of being a chess tournament, it'd be a charade? This isn't about judgement, anyone calling such an abomination a "chess tournament" would be a fool.

What you talk about is called "thematic chess", and people have to be informed before they join about the theme, about the difference from chess, nobody gets to the board and it's surprised that moves they didn't intend to make were forced on their side. This is clearly distinct from chess.

So I'd definitively draw the line at "a player that loves chess and is given the choice between a normal game and X would choose normal chess", because then, this hypothetical player needs to have explained to her the difference between normal chess and X, so X is not chess (under this definition even time handicap games aren't chess, because you need to tell the player about the time handicap which would be different from normal chess with equal clocks, but it doesn't require judgement.)
Chess 960 isn't strictly speaking chess, since the start positions are illegal and the rules slightly different, but it still feels like normal chess.
What about the chess variant where people can promote a pawn normally, or they can promote it to Archbishop (Bishop + Knight) or Chancellor (Rook and a Knight), would you still consider it chess? Because most games without pawn promotions would look exactly like chess, and even for those with pawn promotions, it turns out players wanting to win are most likely to promote to queen anyway, so the extra rules have no effect for those games, because games with an Archbishop or Chancellor appearing would have been decided at that point anyway.

Is your claim that since this feels like chess it isn't fairy chess? Or that it requires judgement? I think this is clearly a chess variant, because players would need to be told in advance about the difference, so one of them isn't shocked when their opponent promotes to an upside down rook that can jump like a Knight or do Rook moves... I can only imagine the face of that person if this was tried to be passed on as chess...

If you need to explain the difference between chess and something else, that something else can't be chess.
Great spirits have always encountered violent opposition from mediocre minds.

Post Reply