Hmm, so it seems there are more potential issues which can arise than just the case @Samraku and I previously discussed about this (bolding mine) :
I just came across another case in which this seems to have backfired and helped a player who was continuing to stall, to somehow win the game without having to play out any sequences.
(They were behind and all of the borders had been sealed, continuing to play inside opponent territory (including moves like the 1-1 point), whilst the opponent who was ahead, passed.
Then the server decision’s AI evaluation changed in their favour after some moves with a follow-up they seemed not to see (and I think would be difficult for most 15k to read out, and easily missed), + a pass, and they instantly claimed victory by server decision.
It seems to not be what the anti-stalling feature was intended for at all to change the victory in a stalling player’s favour based on hypothetical situations some of their moves create, which they didn’t and likely might not see at their level, but which the AI sees.
(and presumably evaluates as certain enough to be above 99% certainty, due to all of the borders being closed).
(actually perhaps rather the opposite of what was intended, as it did somewhat aid the stalling, player and not having to follow-up their otherwise potentially useless moves with working sequences)
(It is true that Black owes a move in the upper left in the game linked after the 2-2 point is played, but it doesn’t seem the anti-stalling giving the victory to White when neither player sees the cut sequence is really what is correct in this case.
At the least, both players should have to agree to the status of that area and/or play it out if they think White can do something there.
If both think it’s dead for White, which seems somewhat likely to be the case given that neither player continued to play there – cut or defend at a huge point, then White shouldn’t win the game.)
Perhaps this should be re-looked into and changed ? @GreenAsJade @anoek