lichess.org
Donate

Lichess Puzzles can get better

As many I've been occasionally trying the puzzles and I must say I find them frustrating. To make my case I will use puzzle number 61578 as an example.
1) This puzzle (As many others) was based on a banned player's move. It doesn't feel good to try and find a cheater's move.
2) Still the above was just something that upsets me, it can still be a good puzzle. The man problem is that (SPOILERS AHEAD) the answer to the puzzle is 1. Qb3 - Rxc5 2. Qb4-R (somewhere) and then 3. Rxh6+ and whites captures the black Queen. This is the variation you must count in order to get in the variation otherwise you just hang the c pawn, right ? Well The Computer answers with Qb3 - e4 (??). I understand that evaluation-wise it might be the better reply but now simply Qxc2 and there is no real chess and/or counting involved. This is an extremely common occurance and it's just bad puzzle solving.

Thank you.
ya its a good rook trap puzzle but i don't simply understand what is the need to play e4
@lovlas The puzzle is very good, the point is that it is being ruined by the way it is being treated. As I said above, I understand that via engine logic "..e4" is better than Rxc5 because it gets black to a +2.5 lost position rather than a +6 lost position, but we are not engines, we are humans. A normal puzzle in a book or even online would have the variation leading to Rxh6+ as the main line.
I've offered suggestions and insufficiently tested patches to attempt to improve the puzzle generator. Many people have suggestions but we aren't generating new puzzles right now and even if we were, we don't have a developer capable of working on the puzzle generator.
If Lichess acknowledges that there is room for improvement it's enough for me. Whenever they can they'll try to fix it

This topic has been archived and can no longer be replied to.