I’m not sure what happened to the scoring in my game. when i pressed “accept removed stones”, the screen looked like this:
then the scoring immediately ended and the screen looked like this:
that isn’t what I agreed to.
Normally what you agreed should be the result (unless your opponent disagreed).
Besides your counting is right as W chain is dead.
Luckily you are still winning (mods can only annul a game if necessary).
To avoid this, i advise to remove the stones by capture (will not affect the score in chinese rules) but that may launch the new anti stalling feature after 3 pass of your opponent. I dunno more.
See for example this:
Is it possible that the opponent toggled the stones a split second before the click that accepted the result (perhaps even honestly thinking it was a seki), and then quickly accepted the result themselves too?
Or more in general, does OGS have any sort of enforced delay or other mechanism between stone marking and acceptance that would help prevent that kind of thing?
But then i guess that the OP would be asked again to confirm the modified position.
There is an delay (something like 1.5 seconds) where the Accept button is disabled any time there is a change.
Actually i think that’s what happens. I have lost games this way also. If true, it’s an exploitable design flaw.
This makes me doubt on score cheating
If so, there’s a problem with the implementation or the delay isn’t long enough.
Cool, yeah that’s what I hoped.
Would be curious to hear about how could have happened then. Either the server has some corner case where it rejects the players’ stone markings and substitutes something else, or perhaps 1.5 seconds isn’t long enough for reaction time if you’re looking elsewhere at the wrong moment and not actively on guard for this kind of race?
Do you remember to have the accept button in grey before you could accept? Then if so did you check if the counting proposal was the same as yours?
I mean i went into this kind of stupid battle and it didn’t seem to be that obvious for my opponent to validate his estimation.
heh. the accept button is supposed to be gray? i guess I’ll pay more attention in future.
It is not grey by default.
It’s grey a short time if your opponent modify. So you can check his modified proposal and not run into a blitzkrieg.
Now how you reported doesn’t seem to be any chance for him to have modified it.
If so, the position can’t be modified in between the time you look and the time you press (will grey and let you wait before it turn back to clear again and show new position on the screen instead)
If you noticed nothing then i suspect a scoring calculator wrong, not your opponent.
Going back years, the scoring clock has been irregular in its function. I have mentioned this several times in different threads. In 2016, when I first joined, the documentation said that the scoring clock had 5 minutes on it and that it would reset every time the board was altered. That was still true when I began moderating in 2018. When AdamR and I and others rewrote the documentation around that time, I verified the 5-minute reset with the developer. However, in the course of dealing promptly with score-cheating reports, I found that the clock often timed out much faster than the supposed 5 minutes after the board was altered. I have long suspected that the scoring clock does not reset as it should after each alteration.
Bottom line: your opponent probably thought the position was seki, and the scoring clock failed to reset.
a bit more info. all of the scoring happened within 1 minute, I’d estimate. i clicked accept on the first board state, and immediately (from my point of view), the final score was set to the second board state.
In that case, it seems certain that the scoring clock (1) failed to reset as it is supposed to do, and (2) failed to run even its initial 5 minutes (invalidating my hypothesis that the initial clock time was still running and ran out). In any case, as I said, I have seen serious irregularities in the scoring clock many times over the years.
Analysing OGS possible failures is quite a process lol.
Scoring clock now…
All this time I’ve been reading that username as “saxman”. Only when I went to search for it was I forced to notice I’ve been misreading 
The timing log of that game shows that saxmaam accepted 4 seconds after the opponent changed the scoring.
I suspect laggy internet could be at play: it seems possible that the button is still enabled on saxmaan’s screen while the update from the opponent is in-flight.
We’ve fiddled with how this works many times: it seems hard to get the right balance between safety, responsiveness and comprehensibility.
(Much like trying to quickly agree to something over a laggy Zoom call 
)
Thanks for looking into that.
Sax Ma’am
That is new info that changes everything. I withdraw my hypothesis about the clock reset.
Do the users just send “I agree” to the server, or “I agree to scoring x”?