Openings Training

I like (thanks !) but in thesee exercises often it's advised the unknown moves of the theory and sometimes normal moves are considered as bad !! Sometimes stupid moves are proposed as truths. We lose fast the won points because " good answers " are unthinkable by an average player and come from the practice of the site and not the chess reality...

I'm a big fan of that training. I have encountered what you describe, yes, but on very very rare occasions only, where the best book moves were not the ones asked for in the exercise.
Overall, that training is really excellent to reach automatisms in the game's first phase!

Yes, this is a known issue. Maybe the opening puzzles have been generated automatically as well?! Many of them don't make much sense.

It would be good to form a team of volunteers, each with good knowledge of one or more openings, and let them create new puzzles with key positions of their openings, so that the "opening training" on Lichess finally deserves its name.

Thank you for your very good answers. Yes I believe that it is a software and not a human being who offers us these positions extracted from the practice of Lichess. Sometimes thus failures. It is often necessary to look for four moves. Two are logical, the third least. Then we try anything (!) to complete the answer and the red wrong appears. We lose fifteen points! Finally the fourth validated move has no logic, absurd ! It once arrives on ten.

Often three moves are sufficients instead of four.... so we have to look for a novice player and sometimes boum ! RED !

Often in the beginning of a game we have two choose between two moves. Three moves is rarely. But in the " openings training " four is normal. Normal ? Why not ten ? The fourth is always a bad and illogical move but not dangerous for his colour. So we don't thinking on it. We prefer a real move and DOWN ! RED !

I understand ! The fourth move is h3 or a6 (always stupid) but the log needs four moves ! How find this ? Berk.........

The opening training should be based on the Masters Opening Book that lichess has instead. If something in there is refuted it'll still have a greater percentage of being right then the current system.

that's actually a good idea

I agree with Tabarjack (thanks for answer !), the database is the reference. However when we understood the moves which the program likes (h3, h6, a6, a3, lol) the resolutions are easier. It is enough to play no risky moves, no compromising moves : program does not like. :-)