asean chess engines and rating list

[hgm]

OK, it is not as bad as KNNK; there are some won positions. Some wins can take as long as 36 moves. But only 0.9% of all positions where the bare King has the move is lost. It seems silly to make a special rule for these exceptional positions.

And the given rules still seem to be self-contradictory, in that they specify a draw after 50 reversible moves. So what is it worth that you can claim draw after 64 moves in KNQK? Am I supposed to understand that the 50-move rule does not apply against a bare King? Or just not in KNQK? Do I get infinite moves KBNK, or just 50?

[Evert]

OK, it is not as bad as KNNK; there are some won positions. Some wins can take as long as 36 moves. But only 0.9% of all positions where the bare King has the move is lost. It seems silly to make a special rule for these exceptional positions.

I suspect that the ending is relatively common, and humans tend to play it very badly (both defending it and trying to win it). I agree you don't need a 64-move rule for it though.

Note that treating the ending as either a general win or a general draw is dangerous. My program penalises the evaluation for it (to avoid trading into it when ahead), but it will sometimes trade into lost positions as a result. Because it uses a tablebase to actually play the position (the tablebase is generated on the fly, but only when needed at the root; I still need to set things up so it will also generate the tablebase if it comes up often enough in the search), it still manages to draw against programs that don't have special knowledge for how to win this ending (pretty much every time, since the King tends to be centralised and so the mate is well beyond the horizon).

And the given rules still seem to be self-contradictory, in that they specify a draw after 50 reversible moves. So what is it worth that you can claim draw after 64 moves in KNQK? Am I supposed to understand that the 50-move rule does not apply against a bare King? Or just not in KNQK? Do I get infinite moves KBNK, or just 50?

I read that as "the general rule is draw after 50-moves", but it gets overridden in the cases listed, so you get 64 moves in KNQK only.

For what it's worth, I think the Makruk rules are superior here: you get at most 64 moves to win but, depending on material, that number may be lower. Counting starts at the number of remaining pieces (which is at least 3).
Interestingly, you don't have to restart the count after a capture (and you wouldn't if it's not to your advantage) so there's an added strategic element there as well. I'm not sure how relevant this rule is in practice; I see it making things more fun in a café or street setting, but for tournament play it seems irrelevant. It may just make the game more complicated with no obvious benefit.

[Evert]

I prefer asean over makruk. Asean has a simplified counting rules, and you can promote to other piece type which is fixed in makruk.

End games are certainly very different, and passed pawns are suddenly a thing to worry about. Whether that makes it a better game than Makruk or not isn't so obvious to me. I can imagine it's less draw-ish, but draws don't have to be boring. Either way, my impression is that it's a good game.
At least people play Makruk though; from the lack of information about ASEAN Chess (in English, but from what I can find with Google Translate, also in Thai) I get the impression that apart from the event in 2011 where the game premiered, no one plays it.

Perhaps 3 months from now i will organize a championships.

I may be able to get something to work in that time frame. Depends on how busy I'll be.

By the way: did you run into the odd "Sjaak hangs" issue with ASEAN, or does that just happen with drop games? I still have no idea what might cause it.

[hgm]

Interesting: if I solve KNFK on a 9x9 board, some 33% of all positions with the bare King to move turn out to be lost. (And 56% of the positions with the strong side to move are won. Note that I count illegal positions where you can capture the King as wins.) This indicates it is a generally won end-game if the Ferz is on the right color, and a draw if it is on the wrong color. (On 9x9 all corners are the same color.) That it is a general draw on 8x8 suggests that it is impossible to drive the King from the safe corner to the lethal one, without it escaping to the other safe corner.

[Ferdy]

And the given rules still seem to be self-contradictory, in that they specify a draw after 50 reversible moves. So what is it worth that you can claim draw after 64 moves in KNQK? Am I supposed to understand that the 50-move rule does not apply against a bare King? Or just not in KNQK? Do I get infinite moves KBNK, or just 50?

As I undestand it, the 50-move draw rule has to be claimed (article 9.2) it can extend if neither player claim it (Article 9.3). In case of bare king like KNQ-K a player can claim a draw via 50-move draw rule instead of bare king counting 64 move limit because bare king move counting limit in article 5.2.e does not mention that it is exempted from the 50-move draw rule in article 9.2.
In article 5.2.e it does not mention about a player claiming a draw, it is the arbiter that will decide that such game is a draw (he may observe the game in progress counting the moves, whatever) and that both players should stop playing .

I sent an email to ncfp - national chess federation of the philippines clarifying these rules, i.e the bare king move counting rules on particular pieces and the relation of the 64 move limit on KNQ-K and the 50-move draw rule.

[Ferdy]

By the way: did you run into the odd "Sjaak hangs" issue with ASEAN, or does that just happen with drop games? I still have no idea what might cause it.

I have not observed a deadlock because winboard would crash.

Also when using command line matching the 2 engines, winboard would crash when Sjaak is involved.

One thing I observe on the Sjaak when run in console is that after typing new and pressing enter, nothing will happen. If I press enter again the # New game 'Chess' will appear.

Code: Select all

new

# New game 'Chess'

Could this be the reason that winboard had already sent the command new and Sjaak has not recognized it yet?

[Ferdy]

I get the impression that apart from the event in 2011 where the game premiered, no one plays it.

There is a tournament in 2013 in Myanmar.

https://www.youtube.com/watch?v=IpePrNvc1VA

https://en.wikipedia.org/wiki/Chess_at_ ... sian_Games

http://www.chess-results.com/tnr118718. ... =30&wi=821

[Evert]

One thing I observe on the Sjaak when run in console is that after typing new and pressing enter, nothing will happen. If I press enter again the # New game 'Chess' will appear.

Code: Select all

new

# New game 'Chess'

Could this be the reason that winboard had already sent the command new and Sjaak has not recognized it yet?

Could you run the following experiment: take the mini-Shogi position where Sjaak stalled (from the other thread), then set Sjaak thinking by feeding it "go". If it stops producing output, hit enter.

One thing I did see is that output from Sjaak would stall and not be printed, even if it was there, as though it was being kept in some sort of buffer. I blamed this on a weird interaction between Wine and console output over an ssh connection, but if it happens in Windows as well it could explain the observed behaviour: Sjaak seems to hang because its output is trapped in a buffer somewhere.

If that is the case, the question becomes why it happens, because all buffering should be switched off at the start of the program (and I tried to get rid of the problem by adding an fflush(stdout) after sending output, which did nothing; this made me conclude it was a problem with Wine).

[Evert]

I get the impression that apart from the event in 2011 where the game premiered, no one plays it.

There is a tournament in 2013 in Myanmar.

Point taken, but that's still 4 years ago. It doesn't convince me that the game is alive and well, unfortunately...

[Ferdy]

I get the impression that apart from the event in 2011 where the game premiered, no one plays it.

There is a tournament in 2013 in Myanmar.

Point taken, but that's still 4 years ago. It doesn't convince me that the game is alive and well, unfortunately...

I am not really concerned if there are less people playing this variant. I like this variant because it is simple and close to chess. Simpler than chess as it has no castle and e.p captures (can be easily learned by starters) and best for human to play against an engine and not get crashed easily .