After having it happen to myself, a slightly possible winning position being timed-out by my opponent I saw this ‘won’ ranked game wasn’t reflected in ‘Wins vs stronger opponents’ I decided to take a look at the forum and found this thread.
A lot of interesting viewpoints (and I certainly subsribe to the notion that except for ‘force majeure’ there are neater ways to end a game prematurely than letting it time out), but I do wonder what now?
I thought the decision was made not to implement the change where games won/lost by timeout don’t affect rank. I won a game by timeout today, but it doesn’t show up at all in my rating history. What gives? A win is a win, a loss is a loss and I don’t think timeouts should have any bearing at all on whether rankings go up and down via correspondence games. This opens the door wide for abuse by people who decided to purposely timeout of correspondence games when they are at a disadvantage and don’t want to lose rank, and unfairly punishes the person who was in a winning position.
6 Likes
I probably missed that decision somehow, but surely can live with it.
One related question: I am playing against an opponent who is likely going to time out for a second time in a row. Is that a reason to report him/her?
1 Like
If timing out doesn’t even count as a loss, it seems highly unlikely that timing out is reportable…
1 Like
I had to go to emergency room and first thing I did was login to all go sites and set vacation, then call 911, then make sure cat had food, water, couldn’t escape. They had to break the door down because I should have unlocked it instead of worrying about go.
3 Likes
It’s a strawman though.
It’s hard to believe that the whole system of ranking should be distorted by timeouts not counting as losses because maybe one day someone might have a real emergency where they couldn’t set vacation.
At the point you’re being carried off to the emergency room, a few games timing out is probably the least of your worries.
(And how many people who play go online won’t have a phone with them in hospital to while away the hours…)
4 Likes
“The whole system of ranking” being “distorted” seems a bit overwrought. I’d bet it wouldn’t have much impact on ratings. Run a simulation or something to find out. What % of rated games of different kinds are decided by time. Frequent changes of that setting could be flagged or limited. If it is effectively being used as a sandbagging tactic, boot the offender off.
I don’t recall any complaints in DGS forums about ratings distortions.
Yeah you know what - you are right - my response was kind of hyperbole. In response to your strawman 
So the basics of it are that the impacts of timeout-not-counting-as-loss are simply that it pisses people off to play a game and have the opponent escape by timing out. That just seems “not right”.
Arguing that it has to be this way because otherwise someone might get timed out when they go to ER just doesn’t seem to cut it.
GaJ
2 Likes
Def: strawman: a person compared to a straw image; a sham. • a sham argument set up to be defeated.
Your reply is out of line.
1 Like
LOL, and whatever.
The basics of it are: the impact of timeout-not-counting-as-loss is that it pisses people off to play a game and have the opponent escape by timing out. That just seems “not right”.
Arguing that it has to be this way because otherwise someone might get timed out when they go to ER just doesn’t seem to cut it.
4 Likes
Maybe it’s just my lack of understanding, but it seems to me that no correspondence timeouts being included would be more damaging to a rating system than all of them being included. As GaJ frequently argues, a loss is a loss.
I feel, especially with this new system, even if a player loses, say, 30 consecutive correspondence games by timeout, their confidence rating would be huge and they could quickly return to their true rank.
3 Likes
The problem with counting all correspondence timeouts is that you have a marvellously convenient tool for sandbaggers. On a plate.
That said, I agree that a result should be a counted result (unless annulled by a mod). I just wanted to point out the harm sandbaggers do. My preference for the K.I.S.S. strategy(Keep It Simple Silly) applies here. I think trying to combat sandbagging is futile. As futile and self-defeating as sandbagging itself.
From a statistical point of view I have no idea which arrangement does more damage to the ranking system but I feel certain that excluding a large category of games definitely does more damage to the credibility and integrity of the site.
1 Like
Sandbaggers can already play a few moves and then resign. This is more time efficient than waiting for games to timeout so I can’t see it being an issue. Furthermore, if we have any kind of punishment for sandbagging, the timeout strategy is just as obvious as the resign strategy.
5 Likes
But all live timeouts are already counted. So if that is a marvelous tool, it’s already there, right?
(What’s more, you can time out much faster from a bunch of live games than from a bunch of correspondence games).
This is only talking about making the rules for correspondence timeouts the same as the rules for live timeouts.
3 Likes
Also, now that they have the new anti-dodger tool… sandbaggers can get a live loss every 2 mins.
1 Like
I honestly think that trying to combat sandbagging with a rules structure is doomed.
There’s something Gödel about it - whatever rule you come up with, will have it’s own doom in it.
Come up with a good rules and practices that work in the absence of sandbagging.
Then monitor behaviour and weed out sandbaggers based on that.
3 Likes
You could do a database search on the glicko ratings for accounts that have sudden swings in confidence ratings. Sandbaggers by definition need to suddenly lose games they should win and suddenly win games they should lose. This will show up in ratings uncertainty which glicko measures.
2 Likes
Undecidability, not inconsistency.
Gödel’s Proof by Ernest Nagel and James R. Newman, 1958, rev. ed. (Douglas Hofstadter) NYU Press 2001.
1 Like
The element of Gödel’s theorem that I was referring to is that any system of rules has the root of it’s own failure in it (where “undecidable proposition” amounts to “failure”).
My comparison was by way of analogy, not direct equivalence.
In the case of “rules to prevent sandbagging”, my assertion is that whatever set of rules you come up with will have a sandbagging solution built into them. It appears this way because fundamentally sandbaggers are doing what normal players do. So we need to look “outside” the system, in the way that jgk suggested, for a solution.
3 Likes
Gödel
1 Like