Big Dog has it right.
"Who decides when the game is lost."
As for examples of won games being chucked...I have thousands on here.
I've posted lads resigning in won positions!
If you feel annoyed if a lad is playing on then the worst thing you can do is tell them.
You are almost certainly going to get them playing on and on and on....
(I never have nor will, then again a lad playing does not annoy me.)
And we must not forget games like this.
Game 9645063
Black appears not to have known how mate with Bishop and Knight.
The game is drawn on the 50 move rule with Black two moves away from giving mate.
So play on, you never know, especially on here where the winning player
can have 30-40 other games on the go (and is annoyed because you are playing on.)
Checkmate is the aim of the game and on here that is followed to the rule
with many games going the full distance after 80+ moves.
I think 80+ gives a good indication of people playing on because they want to.
By then middle game attacks are over and we are in endings.
Lost endings rarely go to checkmate. Once a lad a sees pawn that cannot
be stopped the game is usually resigned. (in OTB chess.)
RHP games ending in checkmate after 80 moves from my RHP DB of 3½ million games..
7,519 games have ended in checkmate.
On a 5 million game OTB database (Standard issue from ChessBase I½ million more games than RHP)
2,801 games have ended in checkmate.
That number I thought was is rather high in one v one chess, (as opposed to
the one v many on here.) A quick scan through the list shows leterally
100's of under 12's to under 8's games included ending in check mate. (Bless them all.)
Also in that ChessBase disc 31 checkmates are given as draws! (I've corrected the above figure.) That is just recording erros.
I'm happy to report that the RHP database has no such errors.