#9 "You should find out if these disconnections are not "random"."
Basically, this is what Lichess does. They count the frequency of a player losing on time, and if it is too frequent, he will get banned temporarily.
However, the actual state of the game (winning/losing) can only be estimated with a high computational effort, and given the high number of games decided by time, cannot be afforded.
#11 not sure if you read the posts above-
I suggested to just count material, not to evaluate the full position using Stockfish.
I am aware that simply counting pawns and pieces is not as accurate as using a chess engine. But if a guy is at -5 points after 20 moves (let say, 30 seconds from the start of the game) and then disconnects (or closes the browser!), then that should be enough for the other player to claim victory.
#12 You might be aware that counting pawns and pieces is not as accurate... other people's opinions may differ:
In the last position, the side with the most "points" is also better according to Stockfish and is also winning (Kb4/Nxa5xb7 and Black's king is first on the other side), so I'm not sure what it was supposed to illustrate. In my opinion, the point of finding a decision rule for rage quits is moot. It is plain obvious when a rage quit occurs, any data scientist with access to the game database will find several criteria that are 99.9+% accurate.
Furthermore, I stick to my "ignore" policy, the community can sort most of the problem out byt itself.
#13 I think you posted that game, because at some point I was up 3-4 points and in the end I still lost the game. That's ok. The opponent didn't disconnect; he kept playing, and he won in the end. That's fine.
But going back to the first example I posted. If I am in a winning position (say, 4/6 points) and the guy GOES AWAY just after having played 10-20 moves, then I should be able to claim victory.
Maybe the disconnection issue is not really a problem if you play against good player, or faster games. But in the 3+2 games and up, it is REALLY annoying when the opponent just hangs up.
There are a ton of positions where the player with material disadavantage is way ahead in the game.
#14 I posted that game precisely to get this sort of feedback and redirect conversation to "what is a won (or lost) position". The community already raises hell when a player capable of checkmate wins "but it wasn't forced" - they'll react even worse when "the player with more pieces" in a lost position gets an undeserved draw.
#15 Of course if the opponent goes away, you should be able to claim victory. Is that not already possible (with a small delay to verify that the opponent has actually gone away)?
#16 I am aware of that. You can have material disadvantage and still be in a winning position.
I see you have an high rating- so probably you have a different experience about rage-quitting.
I think rage-quitters are more common at lower levels, eg. 1600 points and below.
These "casual" players get frustrated much more easily than serious players-
simply because a weak player commits much more blunders.
I don't think you will find many rage-quitters at higher levels.
They are, though, VERY common at circa 1500 (my level..).
And due to these rage-quitters, slow games become unplayable.
As others have said, it should be quite easy to single out rage-quitters based on history; patterns, etc.
And (to repeat), discouraging rage-quitting could improve the game quality on this site.
Sure. But it's not as easy as just counting pieces or points.
#17 the "claim victory" option wasn't displayed, even after 2 minutes the other guy went away- his game time simply went off.
