Even if true, the scenario you called a plausible scenario would not explain the following observation: when I have seen "Lichess is offline" then there is a rating adjustment 0% of the time, and when I did not see "Lichess is offline" there is a rating adjustment 100% of the time.
This is the behavior I'm seeing. Do we agree that this is an undesirable user experience that lichess provides?
About the plausible scenario: if puzzle history is checked only when determining which puzzle to use, then yes, one could cheat by solving the puzzle in another window, after finding out which puzzle is to be solved. But puzzle history could be checked at both puzzle creation time and rating adjustment time.
Even if true, the scenario you called a plausible scenario would not explain the following observation: when I have seen "Lichess is offline" then there is a rating adjustment 0% of the time, and when I did not see "Lichess is offline" there is a rating adjustment 100% of the time.
This is the behavior I'm seeing. Do we agree that this is an undesirable user experience that lichess provides?
About the plausible scenario: if puzzle history is checked *only* when determining which puzzle to use, then yes, one could cheat by solving the puzzle in another window, after finding out which puzzle is to be solved. But puzzle history could be checked at both puzzle creation time and rating adjustment time.
I don't know how hard lichess tries to accommodate frequent disconnectors. It may not be important to them.
I don't know how hard lichess tries to accommodate frequent disconnectors. It may not be important to them.
Do we agree that this is an undesirable user experience that lichess provides?
Do we agree that this is an undesirable user experience that lichess provides?
I think it's mildly undesirable, yes. But your fix may cause more trouble than it's worth.
I think it's mildly undesirable, yes. But your fix may cause more trouble than it's worth.
Good, we agree that's it's not desirable behavior.
Fixing it is a different question.
I am not proposing any particular fix. In fact I don't know how to fix it. I also don't know what causes it. One might speculate that the reason this happens is that when a disconnection happens, some state gets lost (perhaps the state being the connection), and the puzzle then gets marked as seen (equivalent to the user trying a solution), or similar, but I do not know this. Someone is going to have to look at the code for that. As the forum says "bug reports" as one of its topics, and I reported the bug, and there seems to now be agreement that the behavior (bug) actually exists, I have some expectation that some consideration will be given to fixing it.
By the way, I suspect that lichess already checks puzzle history at puzzle creation time, and tries to find a puzzle not in the user's puzzle history.
Good, we agree that's it's not desirable behavior.
Fixing it is a different question.
I am not proposing any particular fix. In fact I don't know how to fix it. I also don't know what causes it. One might speculate that the reason this happens is that when a disconnection happens, some state gets lost (perhaps the state being the connection), and the puzzle then gets marked as seen (equivalent to the user trying a solution), or similar, but I do not know this. Someone is going to have to look at the code for that. As the forum says "bug reports" as one of its topics, and I reported the bug, and there seems to now be agreement that the behavior (bug) actually exists, I have some expectation that some consideration will be given to fixing it.
By the way, I suspect that lichess already checks puzzle history at puzzle creation time, and tries to find a puzzle not in the user's puzzle history.
@mcgoves said ^
I've never seen it do this. Have you?
Occasionally, I solve a puzzle and it just gets a green checkmark or red X (instead of a ratings change). I assume it is due to being solved in the past. That is plausible to me, but I haven't tried to confirm it (it may not be possible to fetch one's entire puzzle history). I don't think it has ever happened while I was offline.
Also occasionally, someone will post to the forum a "hack" they discovered on how to cheat the puzzle system. Some of those involve revealing the solution to a puzzle while offline, then reconnecting and solving it again, for points. I guess that supports your contention that puzzles concluded while offline don't cause a ratings update. The general consensus is to dismiss the problem with the attitude that cheating and boosting your puzzle rating doesn't hurt the rest of us, so we don't care.
The system definitely shouldn't take such tricks into account. After all, the puzzle can always be reconstructed in the editor and analyzed.
@mcgoves said [^](/forum/redirect/post/gaVCNOwV)
> > I've never seen it do this. Have you?
>
> Occasionally, I solve a puzzle and it just gets a green checkmark or red X (instead of a ratings change). I assume it is due to being solved in the past. That is plausible to me, but I haven't tried to confirm it (it may not be possible to fetch one's entire puzzle history). I don't think it has ever happened while I was offline.
>
> Also occasionally, someone will post to the forum a "hack" they discovered on how to cheat the puzzle system. Some of those involve revealing the solution to a puzzle while offline, then reconnecting and solving it again, for points. I guess that supports your contention that puzzles concluded while offline don't cause a ratings update. The general consensus is to dismiss the problem with the attitude that cheating and boosting your puzzle rating doesn't hurt the rest of us, so we don't care.
The system definitely shouldn't take such tricks into account. After all, the puzzle can always be reconstructed in the editor and analyzed.
@OrangeMjau said ^
By the way, I suspect that lichess already checks puzzle history at puzzle creation time, and tries to find a puzzle not in the user's puzzle history.
I doubt it. And I can see why the devs would hesitate to adding an O(log(n)) search to each puzzle offered, if the O(log(n)) search afterward still gets the job done.
@OrangeMjau said [^](/forum/redirect/post/zuK9qKIJ)
> By the way, I suspect that lichess already checks puzzle history at puzzle creation time, and tries to find a puzzle not in the user's puzzle history.
I doubt it. And I can see why the devs would hesitate to adding an O(log(n)) search to each puzzle offered, if the O(log(n)) search afterward still gets the job done.
So what's the plan? Is it going to be fixed?
So what's the plan? Is it going to be fixed?
@OrangeMjau
Just curious, what was the puzzle rated at ?
@OrangeMjau
Just curious, what was the puzzle rated at ?
Plan? I guess it depends on whether a developer sees this thread and decides it's worth pursuing.
Plan? I guess it depends on whether a developer sees this thread and decides it's worth pursuing.