Chess

Rate this contribution

+37

Make the contribution public (lvl 20+)

Comments

26 Comments

Illedan Unnamed CodinGamer

about 3 years ago

✅ Approved the contribution. Comment: Nicely done. Boss seems good enough and engine is robust.

field3 Unnamed CodinGamer

about 3 years ago

✅ Approved the contribution. Comment: It is excellent that random avoids check. The additional rules are also interesting. Thank you for making it.

jacek Unnamed CodinGamer

about 3 years ago

✅ Approved the contribution. Comment: It's good to have a chess game on CG. Made a working bot. Tested draw by agreement. It seems everything works.

reCurse Unnamed CodinGamer

about 3 years ago

Pushed release candidate, added simple wood boss with minimax depth 2 and strict material evaluation. Ready for approval.

PaarthThePro Unnamed CodinGamer

about 3 years ago

It doesn't work. The boss checkmated me but it said it was draw.

reCurse Unnamed CodinGamer

about 3 years ago

Can you share a replay?

Thyl Unnamed CodinGamer

about 3 years ago

why not gradually, 1 pawn league + knight + king then pawn + knight + king + bishop then pawn + knight + king + bishop, turn and finally we finish with the queen ???

reCurse Unnamed CodinGamer

about 3 years ago

Sorry I'm not making 4 additional games for no reason.

skotz Unnamed CodinGamer

about 3 years ago

Nice game. The time limitation will make this quite interesting. The downloadable PGN is a nice touch. Maybe this is just my laziness speaking, but it might be nice if instead of instantly losing upon making an invalid move it just fell back to making a random move. I've seen other games on CG do this, so it's not without precedence. That would help lower the bar for new players so you can see progress without having to implement the more complicated rules first. As it is now you can't even win against the random bot until you spend a decent amount of time building out the rules. Thanks.

reCurse Unnamed CodinGamer

about 3 years ago

I think the barrier to entry has been made low enough. For the problem you mention, you are already given the list of legal moves if you choose to, so then it becomes a simple matter of checking if what you output is in the list, and if not, output random. That should be equivalent to a puzzle of easy difficulty IMO. Thanks for your feedback!

Thyl Unnamed CodinGamer

about 3 years ago

Hello, it's possible a Chess for 4 player, it's really in chess;com

reCurse Unnamed CodinGamer

about 3 years ago

I heavily dislike games with more than 2 players, so I would rather not. Thanks for the suggestion though.

logiqub Unnamed CodinGamer

over 3 years ago

It's really cool, but I didn't test it yet. Just commenting to add that I remember @Di_Masta was working on working on some chess stuff too. After all, he published https://www.codingame.com/... Other than that, the "fear" I have with chess is the amount of engine source code freely available online. It's not really fair competition. Although with 50ms, you need to understand how chess engines work to fine tune it accordingly. Also if it was me, I would change the UCI protocol link to http://wbec-ridderkerk.nl/... I would also definitely include a link to the "Chess programming wiki". (https://www.chessprogramming.org/...). It's probably the best resource for beginners in chess programming.

reCurse Unnamed CodinGamer

about 3 years ago

You're right, chess is very extensively studied, but that's precisely why I thought publishing it was interesting. I see this less as an actual competition and more as an opportunity to experiment and apply all kinds of classical (and modern?) AI theory. However, the extreme constraints of 50 ms on a single-core have also not been well studied IMO, so it could breed some interesting competition on that. For me the competition would have been more on the crazyhouse variant, but it seems rather unpopular so I'll probably only publish the chess960 variant. Good point about linking to the chessprogramming wiki, will add that to the statement. I disagree with the UCI part, as it treats of all kinds of aspects that are irrelevant to this submission, the only relevance is in the move encoding.

Zorg1 Unnamed CodinGamer

over 3 years ago

Choice of input format is nice even if it cannot be managed by the stub generator. A problem I see about the input format choice when reading the statement : Is the order of the different inputs fixed (fen if present, then moves if present, ...) or does it depend of the order of the keyword ?

reCurse Unnamed CodinGamer

over 3 years ago

It depends on the order. Good point, I'll update the statement.

Unnamed CodinGamer

over 3 years ago

Nice contribution :smile:

Illedan Unnamed CodinGamer

over 3 years ago

Great game btw :) Would be nice with more information in the game results ( captures, round end, spawns etc)

Illedan Unnamed CodinGamer

over 3 years ago

Tooltips would be great. Same with some indicator on dead king since the transition to a new game is very fast

jacek Unnamed CodinGamer

over 3 years ago

Nice. You can change title into Chess 960 indicating it's not standard chess. So 1st league is chess960, 2nd league the other chess960 + crazyhouse? Those are quite different games. Maybe consider different multis, because once you beat chess 960, there is no going back. Also visual suggestion: add capture text into timeline.

reCurse Unnamed CodinGamer

over 3 years ago

Yeah I think at this point I'll consider releasing them both separately. Wasn't my intention to have leagues, it's just to show the different options under one entry. Not sure for the different name as I like the simplicity, but I'll think about it. :) I'm afraid of adding more tooltips because it would make it cluttered and lose the separation between both games.

Unnamed CodinGamer

over 3 years ago

Not tested yet, but I really like your idea to let the possibility to choose what inputs the player want. And I remove chess from my endless list of "Games I want to add to CG". ^^

ABOUT US CAREERS PRIVACY FAQ BLOG

The #1 tech hiring platform

SCREENING INTERVIEWING