Forums

Suggesting puzzles to Chess.com?

Sort:
RPetriePlayer1

Does anybody know if you can suggest puzzles to chess.com for anybody to solve?

I thought it could be useful for chess.com because they don't have to keep adding puzzles themselves (they'd just have to edit the suggestions) and it'd be cool to have your own puzzle on chess.com.

Let me know what you think.

Arisktotle

I don't know where you can send them but I am pretty sure chess.com won't need them. Chess.com and Lichess retrieve their own puzzles from databases with games and games played between their members. Even without the latest AI techniques they are very well capable of extracting valid puzzles from sections in those games. All on automatic without human inspection. In a decade or so you will only find PUZZGPT generated puzzles here with a small dedicated Forum remaining for the community of 70+ members insisting on the man-made stuff

RPetriePlayer1
Arisktotle wrote:

I don't know where you can send them but I am pretty sure chess.com won't need them. Chess.com and Lichess retrieve their own puzzles from databases with games and games played between their members. Even without the latest AI techniques they are very well capable of extracting valid puzzles from sections in those games. All on automatic without human inspection. In a decade or so you will only find PUZZGPT generated puzzles here with a small dedicated Forum remaining for the community of 70+ members insisting on the man-made stuff

ok thanks, that makes sense

RPetriePlayer1

Though it would be neat if you could have your puzzles on your account for everyone to see, or something like that

1cbb

I think this could be a good puzzle for chess.com

RPetriePlayer1
1cbb wrote:

I think this could be a good puzzle for chess.com

That's a great puzzle, was it from one of your games? I would never be able to find that over-the-board.

1cbb

I would never be able to find it myself either

Arisktotle
RPetriePlayer1 wrote:

That's a great puzzle, was it from one of your games? I would never be able to find that over-the-board.

This is the problem with puzzles from games. Most are flawed puzzlewise and the engines mining the game databases for puzzles will probably reject 100 good looking candidates before they accept one.

This one is rejected on the basis that there can't be a choice between 2 or more approximately equal candidate moves, i.c. 4. .. Rxe4 and 4. .. Bxe4. Players don't care about having 1 winning line or 10 but solvers have the impossible task to choose one as the Puzzle GUI only blesses one choice.

1cbb

Then we end it after Rxd1

VenemousViper

I think you could just send them a message.

VenemousViper

Also chess.com is trying to get all of the positions that occured in real games and could be used as puzzles. I'm not sure how many they actually get though.

Arisktotle
1cbb wrote:

Then we end it after Rxd1

That's the right answer! And it leads to solvers complaining that the puzzle stops inexplicably and they don't get it. Humans and machines, no match. wink

RPetriePlayer1
SeaWing7000 wrote:

I think you could just send them a message.

Not a bad idea, my only concern is though whether or not they actually respond, because they probably get a lot of spam mail.

My idea is that they could streamline the process by adding a tab under the "Puzzles" tab that you can add your own puzzles, a program will run through the puzzle making sure it meets all the requirements for a good puzzle (such as only one solution, the solution is clearly better than alternatives, etc.), and it gets released for anybody to solve.

simen_s

I just had this position as white to play. I think it qualifies as a puzzle. It's not too difficult, since I found the right move. But forcing.

KnightOfNimue

Yep. Nxe5+. Pretty actually surprise.png (about previous post)

I had this position just now. I would also like a way to suggest it. Except it's not a good puzzle. Only first move is fairly clear. White to move.

I didn't find the first move. I played Nd4. But white has a winning move. It was fun to see the analysis. There is a lot to unpack