Chess324

[CornfedForever]

When will the CRRL start testing on chess324?

Surely this is a joke question??
You probably mean it, but real chess is where it's at...no one plays this stuff.

[Chessqueen]

When will the CRRL start testing on chess324?

Surely this is a joke question??
You probably mean it, but real chess is where it's at...no one plays this stuff.

Back in 2008 before FIDE recognized Chess960 I asked the same question here and received several bad response like this one and look where chess960 is now, everybody is testing it and playing it.

[lkaufman]

When will the CRRL start testing on chess324?

Surely this is a joke question??
You probably mean it, but real chess is where it's at...no one plays this stuff.

Well, the Central Rappahannock Regional Library (CRRL) probably doesn't test chess engines, so "never" is a correct response. But CCRL does test chess960 (FRC), and it seems that here we still see significant increase in engine ratings with each new release. At some point testing groups will have to face the fact that standard chess with good openings is a dead draw with top engines playing at Rapid (non blitz) time controls. They can either go with alternate start positions like FRC (or my chess324), or else stipulate "bad" openings like the unbalanced sets by S. Pohl. It's just a matter of time until this is unavoidable.

[CornfedForever]

When will the CRRL start testing on chess324?

Surely this is a joke question??
You probably mean it, but real chess is where it's at...no one plays this stuff.

Back in 2008 before FIDE recognized Chess960 I asked the same question here and received several bad response like this one and look where chess960 is now, everybody is testing it and playing it.

No, 'everybody' is not playing it. Almost no one really...some exhibitions, but 'pro players' will do anything chess wise for money. I don't see any local or national 960 tourneys...do you? Classical is and will continue to be the main form of chess. People who like to throw engines at one another...well, they have their special problems with real chess.

[hgm]

Hello Michel:

The bishop constellations can be further factored as 3*2 (own color vs opposite color). So we can assign a code [0-2][0-1][0-2] which can be converted to a number if required.

I do not understand your further factoring of 'own colour vs. opposite colour'. Can you explain it, please? Thank you.

If there is a need for assigning numbers to the positions, I would recommend to use the same system as in Chess960. Where you first number the Bishop constellations 0-5, and then add 6 times the relative Queen location (0-2) on the remaining open squares. This can be done for white and black seeparately, and then you can take 18*white + black.

Advantage is that the position is easily reconstructable from the number. With alphabetical ordering of the FENs you might need the whole list for that.

OTOH, I understand HGM idea and was thinking exactly the same. Something like:

Code: Select all

Bishop constellations (B for white pair of bishops and b for black pair of bishops):
0: b+c files.
1: b+g files.
2: c+d files.
3: c+f files.
4: d+g files.
5: f+g files.

------------

Queen (Q for the white queen and q for the black queen):
0: the leftmost possible square (from white POV).
2: the rightmost possible square (from white POV).
1: the remaining square which is neither 0 nor 2 with this code.

------------

Encoding:
N = 18*(3*B + Q) + 3*b + q
0 =< N <= 323

------------

Decoding:
floor(N/18) --> for the subset of 'B' and 'Q'.
N mod 18    --> for the subset of 'b' and 'q'.

B = floor{[floor(N/18)]/3}
Q = [floor(N/18)] mod 3
b = floor[(N mod 18)/3]
q = (N mod 18) mod 3

Hence getting N = 190 for the standard position.

Other less intuitive encoding could be assign b=1, c=2, ..., g=5 (or b=0, c=1, ..., g=4) and sum the files of the bishops to get the bishop constellations, because each sum is unique:

Code: Select all

b+c = 1+2 = 3 (alternative: 0+1 = 1).
b+g = 1+6 = 7 (alternative: 0+5 = 5).
c+d = 2+3 = 5 (alternative: 1+2 = 3).
c+f = 2+4 = 6 (alternative: 1+3 = 4).
d+g = 3+5 = 8 (alternative: 2+4 = 6).
f+g = 4+5 = 9 (alternative: 3+4 = 7).

Then going directly for a second number ({1, 2, 3} or {0, 1, 2}) to indicate the queen position, then repeating for black. Something like 6262 (or 5151) for the standard position, which might be more human readable. The same can be tried with the original encoding of bishop constellations, getting 3131 (using {0, 1, 2} for queens) for the standard position.

Corrections are welcome.

Regards from Spain.

Ajedrecista.

I think the 'own color' remark should be read in the same sense as "the Queen starts on a square of her own color". So after placing King & Rooks each player has 3 open squares of his own color (b-, d- and f-file) and 2 open squares of the opponent's color (c- and g-file). So you could encode the Bishops as c=0, g=1, b=0, d=2, f=4. And then for the Queen's location on the remaining open squares left=0, middle=6, right=12. That would give you the code for one player, and for the total setup you would use 18*white + black.

[Chessqueen]

When will the CRRL start testing on chess324?

Surely this is a joke question??
You probably mean it, but real chess is where it's at...no one plays this stuff.

Back in 2008 before FIDE recognized Chess960 I asked the same question here and received several bad response like this one and look where chess960 is now, everybody is testing it and playing it.

No, 'everybody' is not playing it. Almost no one really...some exhibitions, but 'pro players' will do anything chess wise for money. I don't see any local or national 960 tourneys...do you? Classical is and will continue to be the main form of chess. People who like to throw engines at one another...well, they have their special problems with real chess.

And what are all those tournaments? Why don,t you feed yourself with Corn ?https://en.m.wikipedia.org/wiki/World_C ... ampionship

[CornfedForever]

When will the CRRL start testing on chess324?

Surely this is a joke question??
You probably mean it, but real chess is where it's at...no one plays this stuff.

Back in 2008 before FIDE recognized Chess960 I asked the same question here and received several bad response like this one and look where chess960 is now, everybody is testing it and playing it.

No, 'everybody' is not playing it. Almost no one really...some exhibitions, but 'pro players' will do anything chess wise for money. I don't see any local or national 960 tourneys...do you? Classical is and will continue to be the main form of chess. People who like to throw engines at one another...well, they have their special problems with real chess.

And what are all those tournaments? Why don,t you feed yourself with Corn ?https://en.m.wikipedia.org/wiki/World_C ... ampionship

You clearly did not read what I said. Let me...spell it out a big differently.

On any given weekend in the USA, we have tournaments all over the place - standard chess at different time controls. Heck, I played in one this weekend myself. But where are those 960's?

I don't think you will find it any different on your average weekend in any country in the world. Correct me if I am wrong of course...but I am not.

[lkaufman]

When will the CRRL start testing on chess324?

Surely this is a joke question??
You probably mean it, but real chess is where it's at...no one plays this stuff.

Back in 2008 before FIDE recognized Chess960 I asked the same question here and received several bad response like this one and look where chess960 is now, everybody is testing it and playing it.

No, 'everybody' is not playing it. Almost no one really...some exhibitions, but 'pro players' will do anything chess wise for money. I don't see any local or national 960 tourneys...do you? Classical is and will continue to be the main form of chess. People who like to throw engines at one another...well, they have their special problems with real chess.

And what are all those tournaments? Why don,t you feed yourself with Corn ?https://en.m.wikipedia.org/wiki/World_C ... ampionship

You clearly did not read what I said. Let me...spell it out a big differently.

On any given weekend in the USA, we have tournaments all over the place - standard chess at different time controls. Heck, I played in one this weekend myself. But where are those 960's?

I don't think you will find it any different on your average weekend in any country in the world. Correct me if I am wrong of course...but I am not.

This is pretty accurate, regarding OTB chess, (I did win the U.S. Open FRC championship about 12 years ago, but I haven't heard of another one since), but there is plenty of online FRC/960 play, I find that I can get an FRC game online against a suitable level opponent at almost any time without having to wait very long. Basically FRC is currently used in three scenarios: 1. online play by amateurs 2. Big money events for top pros OTB and 3. engine testing. That doesn't put it on an equal footing with Classical chess, but it does put it on a par with or ahead of many other well known widely played games.

[Time]

Sounds very promising.
Thanks, Larry, definitive this chess variant need to be given a chance.

Speaking of double Fisher random chess (DFRC).

It is very rich in openings count, its also do not request book to differentiate engines.
Main problem is many DFRC openings exceeds the draw range.
I am currently try to clear DFRC from "busted" openings.

chess324 is a subset of DFRC, but has no positions clearly beyond the draw range, and doesn't require the special rules of castling. For some purposes there may be a need for more than the 648 game matches possible with chess324, so there may still be a reason to continue to explore DFRC, but for the vast majority of uses 648 games should be enough between two particular opponents.

I did re-scoring of all 324-chess positions.
Results here:

Thanks. It seems I wasn't quite correct in my judgment that none of the 324 positions would be more favorable for White than the Grob is for Black; it seems that four of the 324 are more favorable for White than 1.g4 is for Black, with another two roughly equivalent to the Grob. I had the right idea that positions with White bishops on b1 and c1 would probably be the best, but perhaps I failed to find the worst setup for Black. So most likely four of the 324 positions are theoretically losing for Black with perfect play, but not by much, so I think the full set remains perfectly viable for engine testing/competition, and certainly for human play. One could always discard those four positions, leaving 320, though this would detract from the elegance of the idea. Anyway, Stockfish could be wrong, or the Grob could be not quite losing. What seems pretty clear is that Stockfish evals below 2.00 in the opening are in general probably not losing; even dividing the evals by 2 seems to overstate the truth. Surely the Grob, bad as it is, isn't quite as bad as just being down an average pawn for nothing, and the SF eval is beyond 2 pawns!

You're welcome!
Keep the full set of this chess variant is quite a good idea (even there is some openings that favors white heavily), in case opening is busted engine should prove that in game play. We anyway waste the time and electricity ...
22/324 (which is very high number) openings have high interest in game outcome result.
P.S. Grob isn't only about pawn losing for nothing, Grob also compromises the white king safety (main reason why Grob is losing).

Impressive that you have already broken the number down to 22. I am currently trying to answer the same question (currently on a i7-7500U laptop). It would be understandable if you want to keep the positions private, should the game ever be popular. On the other hand, they (or some of them) might be an idea for something like viewer submitted openings at TCEC, which can test the waters for (selective) 324 chess. (DFRC was rather successful, after all.) One could hope that the potential for (non-artificial) opening theory exploration in at least some of those 22 positions could be very vast, leaving computers incapable of assessing their theoretical status.

[lkaufman]

Sounds very promising.
Thanks, Larry, definitive this chess variant need to be given a chance.

Speaking of double Fisher random chess (DFRC).

It is very rich in openings count, its also do not request book to differentiate engines.
Main problem is many DFRC openings exceeds the draw range.
I am currently try to clear DFRC from "busted" openings.

chess324 is a subset of DFRC, but has no positions clearly beyond the draw range, and doesn't require the special rules of castling. For some purposes there may be a need for more than the 648 game matches possible with chess324, so there may still be a reason to continue to explore DFRC, but for the vast majority of uses 648 games should be enough between two particular opponents.

I did re-scoring of all 324-chess positions.
Results here:

Thanks. It seems I wasn't quite correct in my judgment that none of the 324 positions would be more favorable for White than the Grob is for Black; it seems that four of the 324 are more favorable for White than 1.g4 is for Black, with another two roughly equivalent to the Grob. I had the right idea that positions with White bishops on b1 and c1 would probably be the best, but perhaps I failed to find the worst setup for Black. So most likely four of the 324 positions are theoretically losing for Black with perfect play, but not by much, so I think the full set remains perfectly viable for engine testing/competition, and certainly for human play. One could always discard those four positions, leaving 320, though this would detract from the elegance of the idea. Anyway, Stockfish could be wrong, or the Grob could be not quite losing. What seems pretty clear is that Stockfish evals below 2.00 in the opening are in general probably not losing; even dividing the evals by 2 seems to overstate the truth. Surely the Grob, bad as it is, isn't quite as bad as just being down an average pawn for nothing, and the SF eval is beyond 2 pawns!

You're welcome!
Keep the full set of this chess variant is quite a good idea (even there is some openings that favors white heavily), in case opening is busted engine should prove that in game play. We anyway waste the time and electricity ...
22/324 (which is very high number) openings have high interest in game outcome result.
P.S. Grob isn't only about pawn losing for nothing, Grob also compromises the white king safety (main reason why Grob is losing).

Impressive that you have already broken the number down to 22. I am currently trying to answer the same question (currently on a i7-7500U laptop). It would be understandable if you want to keep the positions private, should the game ever be popular. On the other hand, they (or some of them) might be an idea for something like viewer submitted openings at TCEC, which can test the waters for (selective) 324 chess. (DFRC was rather successful, after all.) One could hope that the potential for (non-artificial) opening theory exploration in at least some of those 22 positions could be very vast, leaving computers incapable of assessing their theoretical status.

I must have missed something, I don't see anything about 22 positions, what positions, what criteria?