Feedback Puzzles v2

The history view seems to be multi-counting failures a lot. Seems if you get something wrong, then click "view solution" it counts it as getting two things wrong in your history.

Also ELO drop seems to be insane.

Would love to see percentiles added to the ratings and a rating progression graph implemented that shows long term rating changes.

I like how you can choose how easy or hard you want your puzzles.

@Jaytronzero and everyone who's annoyed about rating drops: just do mate in 1 puzzles until you don't get any more points from them, that should bring down your deviation a bit

Would it be possible to "enable" the rating distribution graph for puzzle as well?

- I assume that it should be pretty straightforward as I assume the puzzle rating is stored +/- in the same manner
- It would make the ratings a bit easier to compare ratings (I have the impression that the puzzle rating distribution is wildly different, maybe because lots of new puzzles add points while the player base remains somewhat similar).

I looked at the source code and it seems due to the lack of leaderboard for puzzles ( github.com/ornicar/lila/blob/master/app/controllers/Stat.scala#L10 ).

Not sure if there are reasons for the nonexistance of a puzzle leaderboard (it would be gamed or similar maybe?).
I personally think that it wouldn't hurt (if people have fun gaming a leaderboard, let them have fun).

#1 Games should be timed. Puzzles should not be timed; that's what games are for.

#25 I was going to say lichess.org/blog/X-S6gRUAAGjNX4ki/new-puzzles-are-here explains this, but it doesn't. Puzzles are untimed, so creating a rating distribution or leaderboard would inspire rampant cheating

How does the fact that it's not timed result in cheating? And who would be (negatively) affected by that?
If a few weirdos don't have anything better to do than submit solutions to solved puzzles - let them do that.
The result will be that the top 1% or something like that would consist mostly of cheaters (everyone can look at their profile and figure out that a 3500 puzzle rating doesn't jibe with a blitz rating of 800) and the other 99% have a nice distribution graph.

Also: Cheating is possible in classical or rapid (with dedicated software even blitz/bullet I think) - they still have rankings and nobody is upset if once in a while someone games the leaderboard and undeservingly ends up in a top position.

@oerpli cheating in tactics is way different than cheating in games - the former is just artificially increasing a noncompetitive number on your profile (for this reason there are no leaderboards etc) while the latter is a grave fraud

That's precisely the reason why I think that it doesn't matter that a few people will cheat to get a high score.