Reporting puzzles (weak variations as mainline)

The puzzle training feature (general or by theme) is great. But, in some puzzles, less relevant variations apper as main line
(e.g.: In lichess.org/training/rrMht , after 1... Bg5 2. Qxg5 {instead of the far better 2. Be3} ; I thought continuing with 2... c3 followed by 3... Qa5 would be a good attack, but that line was far more complicated that I thought at first, and involved themes such as vulnerable king, skewer and deflection).

I find similar cases to this in many Lichess puzzles- for me puzzles were always about training patterns - ie; deflect or attract a piece to a square where it can be forked/is pinned/open to a discovery.

In your puzzle many people would see the queen on the kings diagonal and realise they could attack it with the Bishop- however the reply to interpose the Bishop on e3 is such an obvious move to see and so it becomes clear that you will not win the queen via a pin or a skewer and in fact you actually need to calculate about 12 ply of "not exactly forcing moves" before the tactic ends and you win the Bishop or the Rook.

It's personal preference of course but for me puzzles should never be that hard - they should be max 4-5 move maneuvers with a clear intent on each move, a clear theme/themes and end with a capture of a piece or mate.

A lot of Lichess puzzles are like calculating studies - I don't do lichess puzzles to spend 20 mins agonising over a position - I can get a Yusupov book for that.

That will always be a flaw of computer generated puzzles and is also a key difference in the quality of many puzzle books - The positions that are chosen do matter.... just not to a computer.

@userfriendly2 I am okay with hard puzzles, esp. because it is possible to select puzzles here by theme and hardness (rating range). I would love to see puzzles that involve mainly one theme/tactic (chess.stackexchange.com/q/35140/14678 is a related question I asked on stack exchange). Except for very easy ones, many puzzles are mixed-tactic. I am not talking about involving a tactic/theme at a lower level; a deflection puzzle would typically involve a (hopefully simple) fork, pin, etc., so it makes sense to tag it only by deflection theme (otherwise, it might show up when a beginner practices fork). Only issue is, we may not realise it since that theme/tactic is not generally listed as a tactic. Examples include (i) forcing to a square (called coercion is chesstempo.com) [similar to luring to square, better known as decoy/attraction], and (ii) reloader (featured in the excellent book 'Understanding Chess Tactics' by Martin Weteschnik).

But, for many reasons, it is advisable not to practice puzzles that are overly complicated for you (what is complicated is personal & also depends on your current form). I used to do harder puzzles and spend 1/2 or even 1 hour on a hard chess puzzle on a news paper. But, nowadays, I am unable to solve much simpler ones in reasonable time. But, in order to get familiar with the patterns, it is much better to do a large number of relatively easy puzzles.

I didn't know that the puzzles in lichess are computer-generated (are you sure about this?). But, either way, such problems can be solved by involving the community (e.g. by voting down or reporting puzzles). I would love to see the same about puzzles themes (but I admit people esp. beginners very often tag puzzles by wrong themes).

@userfriendly2 "A lot of Lichess puzzles are like calculating studies". Yeah, this frustrates me too. Some puzzles lines here are unlikely to be played by average (or even tournament) chess players (more bot-like or robotic?). This may be due to that the puzzles here are computer-picked.

@CyriacAntony When I say computer generated I mean they are automatically detected/selected from users games by a computer with stockfish. They are not carefully selected by a human chess master because they perfectly convey a theme. This really makes a difference in how well the material can be used to learn for us, the average player.

For me, the best of both worlds is a puzzle book on Chessable or software like CT-art - where the entire bank of puzzles are select by a person.

@CyriacAntony said in #4:
> Some puzzles lines here are unlikely to be played by average (or even tournament) chess players

Well, that's because those lines are good. ;)

@CyriacAntony said in #4:
> @userfriendly2 "A lot of Lichess puzzles are like calculating studies". Yeah, this frustrates me too. Some puzzles lines here are unlikely to be played by average (or even tournament) chess players (more bot-like or robotic?). This may be due to that the puzzles here are computer-picked.

@userfriendly2 said:
> That will always be a flaw of computer generated puzzles and is also a key difference in the quality of many puzzle books - The positions that are chosen do matter.... just not to a computer.

In case you guys don't know, Lichess puzzles are yes generated by an algorithm but they are taken from real games played by real players on Lichess in the past. Chesscom (and probably other sites) puzzles (except puzzles from famous games of the past) are not only generated by a computer but also positions are invented by the program.
The difference consists in the fact that lichess puzzles original positions were at least played once by at least two real players, while there is no certainty about other sites, or at least I don't have it.
I don't know you, but this seems a positive thing to me.

doesnt matter how they got there. The point is to solve the position in front of you for a tactic. We all find ourselves lost in a game and come back with a brilliant puzzle-like tactic in the future. This is what it is training you for. The fact that you didn't see that actually worries me for you as a player. Thanks for the win anyway

@vapePen said in #8:
> doesnt matter how they got there. The point is to solve the position in front of you for a tactic. We all find ourselves lost in a game and come back with a brilliant puzzle-like tactic in the future. This is what it is training you for.

It matters a great deal how they got there - it's why some tactics books are better than others - better material. It's even why some coaches are better than others and why they prize "their material" so much. Some positions are good enough and some aren't regardless of whether their is a tactic there.

And it goes without saying that the point is to solve it.. the issue is how relevant some of the lines are to players and the quality of the positions.