- Blind mode tutorial
lichess.org
Donate

Should there be a way to remove poor puzzles from being shown or a way to skip them?

#1

Some of the puzzles have seemingly incorrect moves. In some of those cases, even if you look at the analysis of the game that the puzzle came from, the puzzle will sometimes have no relation to the game; the puzzle line won't be in the game's line or any alternative suggested lines. And certain puzzle's lines assume your opponent will make a suboptimal move which makes the puzzle rather pointless since you can't know that; most puzzles are about forced moves.

61333 is a good example of this problem (https://lichess.org/training/61333). Spoiler incoming. The puzzle line is: white e5, dxe5, dxe5, Nxe5?, Nxe5 where you're up material because your opponent gave up their knight in the exchange. But you can't know ahead of time that upon e5 your opponent would blunder the exchange so it's not a great puzzle. If you open the analysis and see move 9 you'll notice analysis does say that e5 was the suggested move for white in that position but instead of black exchanging pawns, black follows with Nh5 and keeps its material for the time being, and to allow the bishop to help.

You can see bad puzzles like this reflected in their ratings I suppose, but you don't see the rating until after you've finished. I wish I could see the rating ahead of time and just skip low rated puzzles but this doesn't seem possible in the mobile app (I could definitely be missing something since I'm new). I'm curious if puzzles could be (or are?) shown more or less based on better or worse ratings but it doesn't seem so since I've been shown terribly rated puzzles. It would be nice if poor puzzles could be filtered out or shown less but if that's not possible, even comments would be a nice feature to discuss things like this, or even just to discuss anything about the puzzle, obviously in a section that's not immediately viewable so the puzzle's not given away.

J dunno, I'm just spitballing, but the puzzles seem to need work in how they get shown and how ratings can be used.

Apologies if this has been brought up previously.

Some of the puzzles have seemingly incorrect moves. In some of those cases, even if you look at the analysis of the game that the puzzle came from, the puzzle will sometimes have no relation to the game; the puzzle line won't be in the game's line or any alternative suggested lines. And certain puzzle's lines assume your opponent will make a suboptimal move which makes the puzzle rather pointless since you can't know that; most puzzles are about forced moves. 61333 is a good example of this problem (https://lichess.org/training/61333). Spoiler incoming. The puzzle line is: white e5, dxe5, dxe5, Nxe5?, Nxe5 where you're up material because your opponent gave up their knight in the exchange. But you can't know ahead of time that upon e5 your opponent would blunder the exchange so it's not a great puzzle. If you open the analysis and see move 9 you'll notice analysis does say that e5 was the suggested move for white in that position but instead of black exchanging pawns, black follows with Nh5 and keeps its material for the time being, and to allow the bishop to help. You can see bad puzzles like this reflected in their ratings I suppose, but you don't see the rating until after you've finished. I wish I could see the rating ahead of time and just skip low rated puzzles but this doesn't seem possible in the mobile app (I could definitely be missing something since I'm new). I'm curious if puzzles could be (or are?) shown more or less based on better or worse ratings but it doesn't seem so since I've been shown terribly rated puzzles. It would be nice if poor puzzles could be filtered out or shown less but if that's not possible, even comments would be a nice feature to discuss things like this, or even just to discuss anything about the puzzle, obviously in a section that's not immediately viewable so the puzzle's not given away. J dunno, I'm just spitballing, but the puzzles seem to need work in how they get shown and how ratings can be used. Apologies if this has been brought up previously.
#2

if there were comments, I suppose it would make puzzle ratings for users obsolete since people could cheat although I'm not sure why you'd cheat on a puzzle or if it even matters. I'm not sure the point of having puzzle ratings for users to begin with but that's a separate discussion.

Ratings on puzzles makes sense though but it seems they're just not of much use at the moment.

if there were comments, I suppose it would make puzzle ratings for users obsolete since people could cheat although I'm not sure why you'd cheat on a puzzle or if it even matters. I'm not sure the point of having puzzle ratings for users to begin with but that's a separate discussion. Ratings on puzzles makes sense though but it seems they're just not of much use at the moment.
#3

@spamRoast Can you send a link for a puzzle? I think the answers in the puzzles should be the best move

@spamRoast Can you send a link for a puzzle? I think the answers in the puzzles should be the best move
#4

https://lichess.org/training/61333. My bad, updated original.

Check out analysis. Move 9 white makes a move where analysis suggests e5 is better and would be followed by black Nh5.

Maybe I'm just totally missing something here.

https://lichess.org/training/61333. My bad, updated original. Check out analysis. Move 9 white makes a move where analysis suggests e5 is better and would be followed by black Nh5. Maybe I'm just totally missing something here.
#5

Yes, the puzzle could be made better and 'curated' to delay losing the piece.

This is an issue with puzzle generation and lack of curation. The puzzle is absolutely still true and makes complete sense. The piece is trapped no matter what.

Nh5 can be played, but white plays g4, and the knight has nowhere to go.

This is because black has played Re8 moving the last place for the knight to retreat safely to.

Yes, the puzzle could be made better and 'curated' to delay losing the piece. This is an issue with puzzle generation and lack of curation. The puzzle is absolutely still true and makes complete sense. The piece is trapped no matter what. Nh5 can be played, but white plays g4, and the knight has nowhere to go. This is because black has played Re8 moving the last place for the knight to retreat safely to.
#6

Also the ratings on the puzzles work the same as ratings for players. It isn't set at some rating and it stays there. it changes along with who wins and losses against it, just like a player.

Also the ratings on the puzzles work the same as ratings for players. It isn't set at some rating and it stays there. it changes along with who wins and losses against it, just like a player.
#7

Edit: Apologies, I didn't see your last message. Um, I didn't know that puzzle ratings go up and down based on how they're solved but users can also rate them once they've finished them. I have absolutely played low rated puzzles that were clearly not rated low because they were hard. They were rated low because they were bad puzzles. I'm not the only person who's complained of poor puzzles here and there. Mine here was just a poor example. Back to original response...

I guess you're right. I did see g4 after Nh5 but I suppose my point was that that position can still go into longer alternative lines as seen in the alternative analysis. For purposes of a puzzle I suppose that doesn't matter though.

This turned out to be a bad example but I do believe I've seen other puzzles where there were still problems however, or alternate lines that seemed possible. I have definitely been given puzzles of terribly low ratings too. Some that were just not great puzzles, so I still think some of my points are relevant about how puzzle ratings work, or could work, like if the algo for choosing which puzzle to give a user used weights regarding puzzle ratings more and if there was an option to skip/save a puzzle for later. The same way I could with a puzzle book.

I also wish user puzzle ratings were just not a thing as they can really add a negative element to the experience of personal training. I don't have a rating when I learn openings, study games or positions, or work with puzzles elsewhere.

Anyway, that puzzle was a terrible example of my main point since I was just going down that alternate line and not keeping it simple since all cases involve e5.

Edit: Apologies, I didn't see your last message. Um, I didn't know that puzzle ratings go up and down based on how they're solved but users can also rate them once they've finished them. I have absolutely played low rated puzzles that were clearly not rated low because they were hard. They were rated low because they were bad puzzles. I'm not the only person who's complained of poor puzzles here and there. Mine here was just a poor example. Back to original response... ------------------ I guess you're right. I did see g4 after Nh5 but I suppose my point was that that position can still go into longer alternative lines as seen in the alternative analysis. For purposes of a puzzle I suppose that doesn't matter though. This turned out to be a bad example but I do believe I've seen other puzzles where there were still problems however, or alternate lines that seemed possible. I have definitely been given puzzles of terribly low ratings too. Some that were just not great puzzles, so I still think some of my points are relevant about how puzzle ratings work, or could work, like if the algo for choosing which puzzle to give a user used weights regarding puzzle ratings more and if there was an option to skip/save a puzzle for later. The same way I could with a puzzle book. I also wish user puzzle ratings were just not a thing as they can really add a negative element to the experience of personal training. I don't have a rating when I learn openings, study games or positions, or work with puzzles elsewhere. Anyway, that puzzle was a terrible example of my main point since I was just going down that alternate line and not keeping it simple since all cases involve e5.
#8

#7: Two points I want to reply to:

  • Puzzle ratings are incredibly helpful to have because it allows you to find harder and harder puzzles as you get stronger. It's true that on Lichess some extremely hard puzzles are low rated and vice versa. One issue that causes this is when people guess a natural move without thinking about it. Overall though, ratings are useful and effective at separating puzzles by difficulty.

  • It might seem like you find lots of bad puzzles, but in my experience puzzles on this site have been overwhelmingly good. Definitely do a thorough analysis (like how you did in this thread) before concluding the puzzle is wrong. After having done a relatively large number of puzzles here, I've rarely run into puzzles that were "wrong".

#7: Two points I want to reply to: - Puzzle ratings are incredibly helpful to have because it allows you to find harder and harder puzzles as you get stronger. It's true that on Lichess some extremely hard puzzles are low rated and vice versa. One issue that causes this is when people guess a natural move without thinking about it. Overall though, ratings are useful and effective at separating puzzles by difficulty. - It might seem like you find lots of bad puzzles, but in my experience puzzles on this site have been overwhelmingly good. Definitely do a thorough analysis (like how you did in this thread) before concluding the puzzle is wrong. After having done a relatively large number of puzzles here, I've rarely run into puzzles that were "wrong".
#9

Particularmente, eu tenho feito e gostado bastante dos quebra-cabeças do Lichess. Fiz mais de 700 recentemente e não encontrei 1 sequer com erro. Todos que errei e pensei que pudessem estar quebrados, verifiquei claramente no engine do site que eles estavam corretos e minha ideia de jogada era ruim.

O sistema de rating dos problemas também achei excelente, funciona muito bem. Esse tipo de autoregulação funciona muito bem, a dificuldade de um exercício é aquela lá de fato, é a dificuldade que as pessoas têm. Se você errou um excercício com rating bem menor que o seu, pode ser desatenção ou uma deficiência mesmo. Se fossem ranqueados por um "dono" (como ocorre nos livros de táticas) certamente a divergência seria bem superior.

Apenas compartilho a crítica quanto aos quebra-cabeças com muitos dislikes que ainda estão na fila. Deveriam tirá-los. São poucos, mas já fiz exercício com -3000 de avaliação!

Outros pontos que poderiam melhorar:

  • separação por temas.
  • possibilidade de seleção de quebra-cabeças por ranting (ex. eu quero resolver problemas apenas de rating 1700-1900).
  • quebra-cabeças em abertura, meio-jogo e final de jogo.
Particularmente, eu tenho feito e gostado bastante dos quebra-cabeças do Lichess. Fiz mais de 700 recentemente e não encontrei 1 sequer com erro. Todos que errei e pensei que pudessem estar quebrados, verifiquei claramente no engine do site que eles estavam corretos e minha ideia de jogada era ruim. O sistema de rating dos problemas também achei excelente, funciona muito bem. Esse tipo de autoregulação funciona muito bem, a dificuldade de um exercício é aquela lá de fato, é a dificuldade que as pessoas têm. Se você errou um excercício com rating bem menor que o seu, pode ser desatenção ou uma deficiência mesmo. Se fossem ranqueados por um "dono" (como ocorre nos livros de táticas) certamente a divergência seria bem superior. Apenas compartilho a crítica quanto aos quebra-cabeças com muitos dislikes que ainda estão na fila. Deveriam tirá-los. São poucos, mas já fiz exercício com -3000 de avaliação! Outros pontos que poderiam melhorar: - separação por temas. - possibilidade de seleção de quebra-cabeças por ranting (ex. eu quero resolver problemas apenas de rating 1700-1900). - quebra-cabeças em abertura, meio-jogo e final de jogo.

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