Puzzle algorithm change?

@binyaminaap,

I'm not certain the rationale behind a minimum of +5 was made, for any puzzle solved correctly. I believe it had to do with members taking a long time and the older algorithm, at the time, counting off for taking too long (or maybe just +1). That was changed to give +5, no matter how long it took and apparently no matter the rating difference.

Glad I finally found this topic. This is really stupid. If you get one wrong you get like - 25 lol. If you instantly solve it, still only +5. What's wrong with chess.com. Is your goal to make chess as annoying as possible for your users?

There are many topics about puzzles including "+5" and "-25" things

This is typical of chess.com. Well-functioning features are changed 
for the worse, and criticism of them is ignored.

It's not ignored, it just that there's additional work going towards puzzle so the final planned state isn't ready. It's possible there will be additional tweaks after all changes are implemented.

Literally no one wants changes. We just want the previous functioning rated puzzles system to work again like it was.

I know for a fact there has been a lot of negative feedback about the puzzle rating system over the years, and a lot of it in the forums, and I'm certain a lot more as direct site feedback.

The full set of changes is meant to make things better, in the long run. Whether the final planned stages meet that goal or not could be a different story. It's likely some members won't like the final results. It's also likely the system will get added changes as time passes.

what kind of negative feedback?

Primarily that member puzzle ratings are completely out of alignment with player ratings. The first part of the change is exacerbating it for some members but when the work is completed there should be better alignment, as I understand it.

It's a different system, of course it yields different ratings. The same is the case for Fide Elo, USCF, DWZ, Glicko, and so on. Even Rapid, Blitz and Bullet ratings aren't directly translatable on chesscom. And that is ok.

The most important aspect is that ratings in a system itself are coherent. And this is not even closely the case right now. But with users of the same level and setting getting differently rated puzzles, comparability isn't possible anymore anyhow.

“Solving puzzles earns you 5 points, while losing deducts 20 points. This scoring system is nonsense.”

If I solve a puzzle +5 point, get one wrong -31. This is not making any sense

The only thing I ever noticed about that were observations instead of complaints. I don't think you could compare game play ratings with puzzle ratings because they are two completely different things. Some people have more skill in games. Some people like me have more skill in puzzles. My puzzle rating only went up by hundreds of points after this change and that's not aligning my puzzle score with my game play at all. I am skeptical that things will get better.

How can we fix this issue? Previously, when you solved a difficult puzzle, you earned points based on how quickly you solved it. Now, every puzzle only gives you 5 points, regardless of difficulty. But if you fail, you lose around 20 points. That’s just unfair.

The higher your rank the more points you lose?

I'm 3800 and I lose 31 points with each miss.

Around 3300 I was losing about 20-25 points.

It seems that no matter how fast you solve the puzzle, you only increase your rating by 5 points now. Is this correct?

It's still based on ratings and time to solve. I'm not 100% which ratings are being used for point gain but I still got some a little higher than +5 on my last full season a few days ago.

Hi

Every word + must add:

Not even close to the case right now - and as long as you continue to mutilate Glicko with the +5 minimum for every puzzle, it cannot possibly and will never be coherent, only flawed and illogical.

Please, do all of us a favor: whatever the long-term development plans are, cut the +5 immediately.