Which rules does lichess follow to determine the result of a game if a player is out of time?

The FAQ says that a draw can be declared if an opponent cannot win by any sequence of moves, but still it seems that sometimes lichess declares a win in such cases (lichess.org/Cc2ChRRbesCZ). What are the particular rules that lichess adopts regarding the results of the games that ended with a forfeit of one of the players?

@slowpoke28 said in #1:
> The FAQ says that a draw can be declared if an opponent cannot win by any sequence of moves, but still it seems that sometimes lichess declares a win in such cases (lichess.org/Cc2ChRRbesCZ). What are the particular rules that lichess adopts regarding the results of the games that ended with a forfeit of one of the players?

That's a neat example. I'd say that's a bug in Lichess if it doesn't agree with the FAQ.

That looks like a case where an appeal ought to be successful. The game should have been drawn.

I'm not sure, but I think it's pretty hard to program the rules correctly and there may be bugs in extreme cases like the one you highlight.

EDIT: cross-posted with @AsDaGo who says much the same thing.

> The FAQ says that a draw can be declared if an opponent cannot win by any sequence of moves

why did you stop reading there? the next sentence in the faq:

> In rare cases this can be difficult to decide automatically (forced lines, fortresses). By default we always side with the player who did not run out of time.

also: github.com/lichess-org/lila/issues/9249

#1 its happen before lichess.org/pHeKcmmM
lichess has not found a way to solve it you can appeal the result

maybe we can think of an easy way to do it

maybe if there is only 1 move and after which black/white only has a king left its declared a draw

@for_cryingout_loud said in #5:
> you can appeal the result

they can appeal the result – but nothing will happen. guaranteed. in the issue i linked above, miguel ambrona has a file of all the ~113k games with wrong result according to fide rules that ever happened on lichess until march – and not a single result was changed.

@glbert said in #4:
> why did you stop reading there? the next sentence in the faq:
>
>> In rare cases this can be difficult to decide automatically (forced lines, fortresses). By default we always side with the player who did not run out of time.

Still, this specific type of draw wouldn't be too hard to detect (just check if all legal moves lead to an immediate stalemate).

Engines can't detect fortresses. There's no simple solution.

@Cedur216 said in #8:
> Engines can't detect fortresses. There's no simple solution.

But this game wasn't a fortress, it was simply stalemate on the next move no matter what. That's not hard to detect (you don't even need an engine).

@AsDaGo said in #7:
> Still, this specific type of draw wouldn't be too hard to detect (just check if all legal moves lead to an immediate stalemate).

yes, this specific type - a forced line that ends on the next move. some fortresses are easy to detect as well. but how many games do you reckon have the wrong result because of a forced line that ends on the next move, per unit of time? it's not many, that much i can tell you. and it makes no sense to increase the complexity of the code to detect edge cases amongst the edge cases.

if you took the liberty of actually checking out the github issue i linked, you'd see that there is some effort to port miguel ambrona's code that can detect *all* fortresses and forced lines to lichess. that would make way more sense, because it would detect all the edge cases, not only some rare ones.