Move threshold for annulment of ranked status

[Feijoa]

@Animiral

Therefore such short games should definitely not be counted towards ranking, right?

Yes! Everyone else seems to be talking about using rating as an incentive to prevent escaping, but I thought rating was not supposed to be a reward.

[Conrad_Melville]

Another drawback to increasing the threshold is that rank manipulators, who typically resign after two moves in order to decrease their rank for sandbagging later, will play more moves to reach the threshold before resigning. This will waste more time for the innocent player if those games are manually annulled because they were thrown as part of a sandbagging scheme.

[txwolf]

The rule only applies to correspondence games, correct? Then I don’t care, I only play live games.

In live games though, I wish if players have already played a move, they need to be held responsible. I have quite preparation routines to get ready for a live game:

1. confirm I have indeed a free time window to play
2. find or setup a game
3. get a cold diet coke as soon as game is confirmed
   …

I suspect others do too. So if you quit after 4 or 6 or whatever number of moves without any consequence, I don’t like it. At least I feel the system should discourage this kind of behavior.

As far as rating is concerned, I think it’s less of a factor. Lots of things cause inaccurate rating, for example, resign prematurely. I had quite some lessons myself and don’t resign unless I am absolutely sure I am losing now. Ratings eventually will be fair as long as we keep playing.

[shinuito]

@txwolf

The rule only applies to correspondence games, correct? Then I don’t care, I only play live games.

Why do you assume that?

Also not the greatest entrance to a discussion: “Oh what’s that you’re all talking about? Oh it doesn’t apply to me I don’t care it so, but here listen to what I think.”

Anyway how exactly would one design an automated system to decide when a game should be annulled or not when the following cases exist?

@Conrad_Melville

Another drawback to increasing the threshold is that rank manipulators, who typically resign after two moves in order to decrease their rank for sandbagging later,

@Conrad_Melville

The trouble with a high number is that players see they are at a disadvantage and they cancel—thereby avoiding a ranked loss.

@mark5000

they had a string of opponents who resigned very early, not because they were losing, but because they didn’t want to spend half an hour against a player who didn’t challenge them at all.

@yebellz

Suppose, I start an even game and I get black. If I play at 4-4 and my opponent immediately responds with a 3-3 invasion underneath, which I don’t like, then is it acceptable for me to quit that game without rating penalty?

@Vsotvep

Except, now that we’re at this topic anyways, I’d like to see that players who regularly resign a lot of games while being far ahead according to the analysis

@Vsotvep

Black has been ahead with 20+ points for the last 50 moves and the estimated win percentage for Black is 99.9%.

I don’t see a problem with auto-flagging (of course if I’m a mod maybe I don’t want every game flagged) but auto-annulling I expect will cause issues and complaints.

• If you blanket auto-annul games with less than N moves but N is bigger than 2, then you give an player the option to choose whether they dislike the current opening or the progression (win/loss) and then they can escape.
• If you auto-annul based on winrate or score with less than N moves, then you need to do a check whether the game is handicap or not since that completely skews things.
• If you auto-annul when the winning player resigns in less than N moves, it’s probably completely dependent on the game whether it’s a legitimate resign or not: It could be a player receiving a handicap thinking they’ve no chance but they still do; it could be a player misreading something, not realising a group could be saved\killed and resigning based on information they have, not what katago has. (I guess this is unlikely in games with ~20 moves, but plausible further into a game.)
• If you auto-annul when the losing player resigns in less than N moves, it could be the case that the losing player blundered and a corner died putting them like 20 points behind. I think this could be a legitimate reason to resign and not have the game auto annulled (like falling into an opening trap in chess, and not losing but being down material and not wanting to play on.)
• What about fast time setting games as well? Some people play very fast time settings, (even compared to me finding 10s blitz fast) and I imagine they will not like if many times an opponent times out it’s annulled. I can imagine people unused to say 10s (or less) blitz timing out if the first complicated fight happens before say 20 moves.

I don’t think auto-annulling purely based on the move number has a chance of working, and I don’t think just increasing the move number is a good idea as others mentioned.

[Platybelodon]

In what universe does someone accept a game, play a move, and then quit in good faith? A child taking your phone? We’re talking about 5+ moves like it’s expected but it’s frankly not at all.

Why isn’t accepting a ranked game a sufficient threshold? Do we have a lot of people complaining about accidentally accepting ranked games? Because that’s a different problem that would need to be addressed elsewhere in the UX.

Edit: I’m imagining someone signing up for an over the board tournament and not showing, and then complaining that they lost rank or lost the reg fee. I’ve yet to be convinced that accepting a game is easy to do accidentally.

[txwolf]

@shinuito

Why do you assume that?

Don’t lecture me when your entry of discuss starts with a question

[mark5000]

@Platybelodon

In what universe does someone accept a game, play a move, and then quit in good faith?

The game I shared tries to supply an example of this. Suppose I make an open challenge with moderate time settings, restricting rank to my level. My intent is to quickly get a serious match against a player my level. Someone accepts. The system shows that they’re my level, and it let them take my challenge, so I’m gearing up for a good game. They open on the side. “Weird opening,” I think to myself. I respond in a corner. Then they attach to their own stone for move 3. At this point, I realize that this isn’t a good match, my opponent is either new or trolling, and I’m stuck playing this game for the next 30-60 minutes if they use all their time. Unfortunately, I set aside time for only one good game, so I decide to take the rank hit and try again.

Who could fault such a player? I couldn’t. But if the general feeling is that the benefits of annulling these games automatically does not outweigh the drawbacks, then there’s nothing more to discuss.

[Platybelodon]

I think this ties back to the subjectivity of what constitutes a good opening, what throwing looks like, etc. Because “we know it when we see it” and I get how frustrating that is, but implementing some kind of rule for this is near intractable. Weird openings have to be protected.

[Platybelodon]

In other words, you don’t really decide what a good game is. Your opponent is trying to disrupt you constantly, within and outside of the boundaries of generally accepted definitions of Good Play

[teapoweredrobot]

@mark5000

if the general feeling is that the benefits of annulling these games automatically does not outweigh the drawbacks, then there’s nothing more to discuss.

I think the solution is possibly worse than the cure. Surely the answer would be to manually adjust the player’s rank so that it actually reflected their rank (or find a way of setting it correctly in the first place) rather than tinkering with game settings and whatnot which firstly doesn’t help with the problem (that player’s rank will remain unchanged having just played an annulled game) and secondly enables shenanigans which we generally expend effort trying to avoid.

@Platybelodon

In what universe does someone accept a game, play a move, and then quit in good faith?

And on this. The scenario I always think of is that you quickly click to accept a game because of you don’t you miss the boat and only then check out the settings/your opponent, see something you don’t like and cancel it/resign. Seems ok to me since if you checked out the opponent/details of game settings before accepting you wouldn’t get the chance to accept!

[Feijoa]

@Conrad_Melville

Another drawback to increasing the threshold is that rank manipulators, who typically resign after two moves in order to decrease their rank for sandbagging later, will play more moves to reach the threshold before resigning.

So if you want to make sandbagging more efficient, why not just offer two buttons for the first few moves, “cancel game” and “sandbag”? That would make it easier to auto-flag the sandbaggers, too.

It sounds like people are used to browsing their opponents’ profile pages to see if there’s an issue, so how about not making it so hard to interpret a player’s game history? For example, if I’m looking for annulled games:

To figure out what happened in those two, I have to compare the four shades of gray on the left to the cut-off, crossed-out text on the right. It seems like we trying to obscure the results. Why not at least show it like this:

To make it better, show the number of moves played or even board images so we can look for unusual behavior without clicking into each game. Don’t show dozens of correspondence games above the more important game history. Make it not scroll all the way back to the top after clicking into a game and going back. Put a warning flag in the player box if there is suspicious history to encourage visiting the profile page.

I had another idea for profile pages: how about a feed of recent game chat messages? This would make a lot of bad behavior immediately obvious and also have side benefits, like giving people a chance to notice comments and reviews posted to games long after they have left.

[stone.defender]

@pdg137

Put a warning flag in the player box if there is suspicious history

what is suspicious history? Its road to existing of discriminated layer of society.