Can Engines properly assess openings? - Chess Forums - Page 6

Optimissed

Jun 22, 2022

0

#101

HolyCrusader5 wrote:

Can an engine determine an opening is sound from the perspective of a "perfect player" or do they lack the intuition and long term planning to do so? I am asking purely out of curiosity.

No, of course not.

MaetsNori

Jun 22, 2022

0

#102

HolyCrusader5 wrote:

Can an engine determine an opening is sound from the perspective of a "perfect player" or do they lack the intuition and long term planning to do so? I am asking purely out of curiosity.

Today's top grandmasters often disregard the top engine moves, for various reasons.

Sometimes, the time engine moves simply lead to forced draws.

Often, the less accurate engine moves lead to better (more practical) winning chances, since they keep things more complex, less familiar, and/or less clear.

Caruana once said that he sometimes chooses slightly "worse" opening lines, simply because they may keep things unclear, and can give the opponent more chances to go wrong.

Conclusion: the top engine assessments, in the opening, are not always the best way to play.

NM athlblue

Jun 22, 2022

0

#103

Also, as preparation, even if the engine refutes lets say one move because there is only one possible answer, and all other options are bad, you can play that move as a surprise because human is not engine. So you have to search beyond computer analysis for good opening preparation.

MARattigan

Jun 23, 2022

0

#104

ogpu-jd wrote:

MARattigan wrote:

drmrboss wrote:

tmkroll wrote:

There's a line in the Traxler where a lot of people on this forum could play better than Stockfish. We were debating it here a few years back. Stockfish says White is winning until it sees black has a draw by repetition, then its evaluation goes to 0. Stockfish will take that draw but people who read that forum would castle Queenside as black. Eventually Stockfish sees black is better but it takes it a very long time. There's a line in the KG that at least five or six years back Fritz was similar, idk about now. Of course engines will never play into either of these lines if you don't make it do it because their opening books have been programmed by human players who have studied and know they are bad.

Which version of Stockfish you use and how many nodes SF searched for that position?

Do you mean analysis by this crappy chess.com server stockfish? In fact if SF search only a few hundreds nodes per move, her strength will be like 1200, but a few hundred million nodes per move will make her like 3500.

Show me the position, and I will analyse in 3 mins and show you how strong stockfish is. ( Let me see whether SF really played bad)

Try your SF out on this position. It's a well known win for White, but my SF can't play it for toffee.

As mentioned in a previous thread my SF evaluates the following position, which you yourself posted, as +6.40 no matter how long I leave it running, whereas Black has a very easy draw.

This one even gets +7.34.

Throw in an extra piece and it does no better. It evaluates this win for White at 0.16 at depth 30 both before and after it blows it on its second move.

In fact endgames tend to get more complicated the more men there are on the board. The maximum length forced mates with perfect play (no 50 move rule) are something like 28, 43, 127, 262 and 594 for 3,4,5,6 and 7 men respectively.

SF appears to play and evaluate 3 man positions perfectly (if you take +ve, 0 and -ve evaluations to mean wins draws and losses) but it already starts going awry with both evaluation and play with 4 pieces (it can't play KBNK accurately). With 5 or 6 pieces it starts losing half points.

Can you really believe that in spite of that, when it gets up to a 32 piece endgame, it starts to give accurate evaluations?

And what do the evaluations mean anyway - there's nothing in my SF documentation that tells me. Positions after all are either won for one side or drawn; there's nothing in between.

I know that im about 3 years late but i guess it could be interesting to see how modern SF15 would do with the presented problem:

For the first given picture: if we assume the FEN to be {7k/N7/p3N1K1/8/8/8/8/8 w - - 0 1} SF finds Mate in 4 in <5s; if we assume the FEN to be {7k/N7/p3N1K1/8/8/8/8/8 b - - 0 1} SF finds mate in 4 in <1s w/ a intel 9600k

The FEN for the first picture is 8/8/8/8/8/1K1N3p/7N/k7 w - - 0 1 (i.e. none of the above). You can see the FEN by clicking in on the two fingers icon below the diagram and then on "PGN".

In either case White can mate the black king in the corner he occupies in four moves so long as the pawn doesn't both promote and have a further move. In your diagrams this is possible because Black needs at least five moves to promote the pawn. In the original he needs only two moves after the pawn is unblocked so with White to play it isn't possible. (With Black to play he can incorporate a check en route to mate if the black king doesn't move so with that alteration it would again become possible.)

The upshot is that White must extract the black king from his corner and force mate on the h file, which makes it mate in 44. This is well outside the capability of any version of Stockfish without tablebase access with practical resources. The maximum depth mate by White any of the SF versions I have can manage with this material varies by a few moves according to version and black pawn location but doesn't exceed 36 moves with the pawn on h3 (compared with an average depth for such mates in the Nalimov tablebase of almost exactly 58 moves).

By illustration here is SF15 attempting the original mate on one core of a Pentium J3710 @ 1.60GHz with a hash table size of 2GB and 20 minutes on its clock.

It draws in 3 instead of mating in 44.

The resources are in fact something of a red herring. I had at one time the last version of Rybka with the 'e' suffix in its version number. That was tailored for specific endgames without using tablebases. It was at least fifteen years ago and running on an ancient IBM PC with a specification in Mhz it would have had no problem with the position I posted.

You can't assume all is uniform progress. E.g. here it is again against SF8 with exactly the same set up. (The final position is less obviously drawn, but my king is in Troitzky's drawing zone with the pawn blockaded on h3 - see the second diagram with the 'X's here.)

It draws in 6 instead of mating in 44.

It takes me twice as long to draw against SF8 as it does against SF15.

ogpu-jd

Jun 23, 2022

0

#105

MARattigan wrote:

ogpu-jd wrote:

MARattigan wrote:

drmrboss wrote:

tmkroll wrote:

There's a line in the Traxler where a lot of people on this forum could play better than Stockfish. We were debating it here a few years back. Stockfish says White is winning until it sees black has a draw by repetition, then its evaluation goes to 0. Stockfish will take that draw but people who read that forum would castle Queenside as black. Eventually Stockfish sees black is better but it takes it a very long time. There's a line in the KG that at least five or six years back Fritz was similar, idk about now. Of course engines will never play into either of these lines if you don't make it do it because their opening books have been programmed by human players who have studied and know they are bad.

Which version of Stockfish you use and how many nodes SF searched for that position?

Do you mean analysis by this crappy chess.com server stockfish? In fact if SF search only a few hundreds nodes per move, her strength will be like 1200, but a few hundred million nodes per move will make her like 3500.

Show me the position, and I will analyse in 3 mins and show you how strong stockfish is. ( Let me see whether SF really played bad)

Try your SF out on this position. It's a well known win for White, but my SF can't play it for toffee.

As mentioned in a previous thread my SF evaluates the following position, which you yourself posted, as +6.40 no matter how long I leave it running, whereas Black has a very easy draw.

This one even gets +7.34.

Throw in an extra piece and it does no better. It evaluates this win for White at 0.16 at depth 30 both before and after it blows it on its second move.

In fact endgames tend to get more complicated the more men there are on the board. The maximum length forced mates with perfect play (no 50 move rule) are something like 28, 43, 127, 262 and 594 for 3,4,5,6 and 7 men respectively.

SF appears to play and evaluate 3 man positions perfectly (if you take +ve, 0 and -ve evaluations to mean wins draws and losses) but it already starts going awry with both evaluation and play with 4 pieces (it can't play KBNK accurately). With 5 or 6 pieces it starts losing half points.

Can you really believe that in spite of that, when it gets up to a 32 piece endgame, it starts to give accurate evaluations?

And what do the evaluations mean anyway - there's nothing in my SF documentation that tells me. Positions after all are either won for one side or drawn; there's nothing in between.

I know that im about 3 years late but i guess it could be interesting to see how modern SF15 would do with the presented problem:

For the first given picture: if we assume the FEN to be {7k/N7/p3N1K1/8/8/8/8/8 w - - 0 1} SF finds Mate in 4 in <5s; if we assume the FEN to be {7k/N7/p3N1K1/8/8/8/8/8 b - - 0 1} SF finds mate in 4 in <1s w/ a intel 9600k

The FEN for the first picture is 8/8/8/8/8/1K1N3p/7N/k7 w - - 0 1 (i.e. none of the above). You can see the FEN by clicking in on the two fingers icon below the diagram and then on "PGN*.

In either case White can mate the black king in the corner he occupies in four moves so long as the pawn doesn't both promote and have a further move. In your diagrams this is possible because Black needs at least five moves to promote the pawn. In the original he needs only two moves after the pawn is unblocked so with White to play it isn't possible. (With Black to play he can incorporate a check en route to mate if the black king doesn't move so with that alteration it would again become possible.)

The upshot is that White must extract the black king from his corner and force mate on the h file, which makes it mate in 44. This is well outside the capability of any version of Stockfish without tablebase access with practical resources. The maximum depth mate by White any of the SF versions I have can manage with this material varies by a few moves according to version and black pawn location but doesn't exceed 36 moves (compared with an average depth for such mates in the Nalimov tablebase of almost exactly 58 moves).

By illustration here is SF15 attempting the original mate on one core of a Pentium J3710 @ 1.60GHz with a hash table size of 2GB and 20 minutes on its clock.

It draws in 3 instead of mating in 44.

The resources are in fact something of a red herring. I had at one time the last version of Rybka with the 'e' suffix in its version number. That was tailored for specific endgames without using tablebases. It was at least fifteen years ago and running on an ancient IBM PC with a specification in Mhz it would have had no problem with the position I posted.

You can't assume all is uniform progress. E.g. here it is again against SF8 with exactly the same set up. (The final position is less obviously drawn, but my king is in Troitzky's drawing zone with the pawn blockaded on h3 - see the second position with the 'X's here.)

It takes me twice as long to draw against SF8 compared with SF15.

if i look for {https://syzygy-tables.info/?fen=8/8/8/8/8/1K1N3p/7N/k7_w_-_-_0_1} its DTZ 84 and not DTM 44, regardless of that: an interesting topic to talk about + i learned something about the "sharing" mechanic

MARattigan

Jun 23, 2022

0

#106

@ogpu-jd

The DTM figure I quoted is in White's moves.

Figures quoted on the syzygy-tables.info site are based on ply, which accounts for the factor of roughly 2 discrepency.

If the site shows DTM figures next to a move you can take that (+1 to make the move) as the distance to mate, in ply, with no 50 move or triple repetition rule in effect. In your link it shows minimum DTM 86 ply after Kc2 so 87 ply altogether or 44 moves by White and 43 by Black.

The DTZ figures do not tell you the distance to mate. They have to be read with care in any case and if the 50 move and triple repetition rules are in effect apply only to ply count 0 positions. They're connected with the distance to either mate or the ply count being reset (but, in particular circumstances which don't affect the outcome of recommended play, can be out by 1 ply) and the figure shown may have 100 added as a flag to show that in some phase of the mate, but not necessarily the current phase, the 50 move rule will be exceeded.

E.g. If you follow this link: https://syzygy-tables.info/?fen=1n1k4/6Q1/5KP1/8/7b/1r6/8/8_w_-_-_0_1 you see no DTM figures so you cannot tell the distance to mate. It shows DTZ of 109, which means in fact there are at least 9 or at least 109 ply to mate or resetting the ply count to a position which is winning without the 50 move rule in effect, but the leading "1" (could be "n") indicates it is not winning with the rule in effect. (You have to work out for yourself or follow the recommended moves to find out if the figure 9 or 109 represents the ply to mate or resetting the ply count in the position you provide - in this case it's 9 ply to reset the ply count.) The DTM without the rule is in fact 549 moves by White or 1097 ply, but not shown on the site.

I didn't understand your reference to "sharing".

Edit: OK, I twigged "sharing". You're talking about the two fingers icon.

gik-tally

Jun 23, 2022

0

#107

[quote]Show me a single opening where you can play better than an engine? [/quote]

in looking at chess games and trying to build my own books, i CONSTANTLY see stockfish recommendations that perform TERRIBLY over the board as played by actual HUMANS. in one particular position SF rated as +20, human players were still LOSING more than winning!

i only use stockfish to verify moves, for the most part, unless its evaluation is at least a full pawn stronger. otherwise, i prefer HUMAN moves. not only will you be learning moves that make sense to actual humans, but you can see how a move performs against another move (it's good to have a sample size of at least 100 games each), and even better, see what to expect in the next move. you waste no time chasing engine or GM moves you'll never actually see over the board. i'm finding AWESOME theory on lichess like that. i just finished my "ice queen" book, and it is winning (I'm talking +3 or MORE) in 99% of the games! only today did i find ANY games that were less than equal. it was just a few lines, and even then, players aren't making stockfish moves and exploiting anyways.

i also hate how wishy washy engines are. they make a lot of "non moves"... pushing for 1/100th of a pawn instead of making something happen. i WISH i had access to junior. THAT engine has an appetite! i've read that leela is better in the openings too where stockfish is better at endgames.

if you ask me, look at a database and write your own theory... picking and chosing moves you like (I REFUSE to ever play pieces back to their starting squares! fritz 6 was HIDEOUS for that! i quit using it to write theory because the free crafty15 that came with it was actually much more aggressive. i could play against fritz into the 30s any day, but crafty would do the job in 20 moves. THAT is my measure of a "good engine".

leela is reported to play much more "human" and it loves gambits. there was a game where it sacked a rook just to get it out of the way before eventually winning.

if you're using them for theory, they'll steer you towards moves humans don't work well with sometimes. they're better for evaluating human moves than suggesting them as far as i'm concerned.