I think there are better suggestions than the streak one.
It seems enough to line up correspondence games so you finish the ones you are winning, playing slow in the ones you are losing (which people do anyhow, so it would not look suspicious) then timeout all the ones you are losing in one big go.
If we’re going to fix it, don’t tweak around the edges.
Not sure if you mean one who is a cheat or one who is innocent, or both. However, an example of an innocent serial timeouter is claudia (player 319815), who timed out of all 20 of her correspondence games, disappearing from OGS. I feared she may have died, but I now suspect that her parents banned her from playing in order to concentrate on her schoolwork. She was/is a fine and gracious player.
I think he means “for all this worry about serial timeout cheating, we have yet to be pointed at one real case. Could someone please point at a real case?”
Which is a fair request, of course. I have had a couple in the past I recall thinking they were likely suspects, but it is a real trawl to try to re-find, so I am waiting to see another suspect case…
Moderator Adam, I cannot fathom any consideration to close this thread when valid ideas are still being offered, no one is being offensive, and no harm is being done by keeping this meaningful subject open.
I think there is some confusion about “serial timeout”. Serial timeout means more than one time out in a row, i.e. in a series. It can be a repeat cheater using the rule to their benefit over and over, but it can be a one time or infrequent “timeouter” who times out a set of games for any reason, such as the person andysif played that timed out of five games at once, or the young lady mentioned above that simply abandoned her twenty games and left the website. Even if a person is playing two games, and legitimately forgets to return to the website before both games time out, the second game is not counted…in essence, it is annulled.
I ask this question: If serial (more than one in a row) Timeouts adversely affect the rating system, then in live or blitz games on OGS, are more than one timeout in a row not counted toward ratings (I.e. annulled by the OGS system)
That is, if I play and time out on two or more live/blitz games in a row on OGS, is only the very first game counted toward ratings?
I haven’t followed this thread … but isn’t any discussion of “serial timeouts” and their effect on rank only relevant for correspondence games?
In the hopes that I understood the question and am not wasting time on something irrelevant 
The whole subsequent timeout rule is relevant only in correspondence games. Timeout itself does not affect anything. And one can hardly time out 20 live games in succession by accident. Live games require both players to be present. And are played one at a time.
The whole problem is (as discussed far above) that in correspondence it is common for players to have 20, 30 or even 50 correspondence games active. And every so often (and it is not uncommon) these players just stop playing, switch to different account, forget to set their vacation or whatever. That’s what causes the whole problem.
Adam, how many successive correspondence games must a player time out in order to adversely affect the ratings system?
Well to be honest ma original reaction was “how the hell am I supposed to know?” 
This really is an unanswerable question (esp. for me). Obviously, any undeserved victory (the timout itself does not yet need to be bad), affects the rating. But that is to be expected, small fluctuations are unpreventable. So the following question would be how much do you consider to be an “adverse effect” and even then it would depend on many factors, like the rank difference between the players, whether the player whose rank droped to 30k because of the timeouts comes back and starts ranking up again and so on. I have no idea, and do not see the relevance. The point is that the dev team sees it as a source of potential problems and while one person timing out of all games might not change much, those mistakes would snowball over time. That’s why the current rule is in effect.
yep
Thank you.
The relevance is that the wins should be credited as ratings “wins” within the mathematical necessity to avoid a statistically significant change in the overall rating system. Example: If ten successive timeouts is enough to mathematically change the overall rating system, then the first nine wins should be rated for everyone, and successive serial timeout wins (>10) should not. It’s very relevant.
Hey Andy,
I don’t think I said or anyone else said that you claimed that you were winning in any games that timed out…?
You said:
Sure I’m not implying that you considered that you were winning in this game but in the context of the discussion on this thread and some other threads I assumed that you too perhaps felt aggrieved that this timeout win didn’t increase your rating.
In fact you did say:
Which I think you mean that you would like timeouts to count. I apologise if I’m way off the mark.
–
The thrust of these discussions appears to be:
Given that you and Tongue appear (to me at least - more apologies if you are not) to generally be more vocal in respect of 2. above, I wanted to see for myself who these serial timeouters are and so I went to your game history.
In my opinion the people in yours and Tongue’s recent game history that have timed out are not trying to game the system by way of the current time out rule. I could talk more at length if you would like about my thoughts on why people time out and specifically in a serial manner. As to whether they should sort out their life priorities so that they pay more attention to their games, that’s a matter of opinion and not really relevant - we’ve got to live with it.
As for allowing timeouts to affect rank, for reasons that have previously been mentioned, the current rule is correct. To change it as some want would require technical fixes - Anoek has plenty more important things to do.
And so, as GreenasJade kindly clarified for me, please can someone point out a real serial timeouter who is gaming the system for their own advantage who actually matters to any of us !
oops! Even more apologies are in order - sorry I timed out in a game last summer against you Andy!
I’ve come to realise that it ultimately doesn’t matter if the person is trying to game the system or not.
It is much worse if they are, but it is still an affront to have been winning a match and not be ranked for it due to the person timing out for any reason.
GaJ
OrbitalEccentric, in this thread my post has a Timeout offender that “actually mattered to me”. A cheater. I'm fed up
A quick look at your history, Orbital, shows
https://online-go.com/player/490644/
Or
https://online-go.com/player/479382/
Or (check out how many of this guy’s games were voided!).
Not serial cheaters (maybe player unstable is, though) but just as an examples, your recent “Win by Timeout” didn’t count for these opponents, OrbitalEccentric. And one of yours after 208 moves! This is happening to you all the time❗️ I’m only looking back a couple days! In fact, i believe that MOST of your Wins by Timeout are unranked, OrbitalEccentric. The majority were unranked for you and you don’t even know it.
It’s been proven many times, so we now go back to solutions. And, before someone says “Look how much that bunch of Tineouts would have swung the ratings system if they counted”, the counterargument is that a) the rating system is swung by them not counting also, and, as b) it allows players to cheat, it swings it even more than if cheating was eliminated, and c) since people who “Win by Timeout” don’t even know their game was unranked, they continue to play games with the “Recenlty Timed Out of a Game” player, turning off that badge, thus negating the benefit of the badge for all others (and allowing cheaters to continue to play new and old opponents indefinitely).
I think these games should say “Annulled By Timeout”, start with that small change.
Anyone else…brainstorm ideas to help our team solve this. Thanks for all input.
I totally agree with this proposal.
Having a game not count and not show up as such is counterproductive at best.
You mean the player whose stats look like this?
Looks rather clean to me. If anything, he’s the one being serial-timeouted. 
All these talks about affecting the ranking system, had anyone actually done any analysis to quantify the effects?
I have gone through my record for about 2 months yesterday, (don’t have the file here, just recall from memory), and have gathered the followings:
I am about 11k, have faced opponents from 18k to 6k.
Any win will score anything from 5 pts to 35pts, depending of course on the rank of the opponent.
We can assume that any “undue win” is when a stronger opponent timeouts, so let say it gives you about 30 pts and bumps you up 2k. Let say from 11k to 9k
After the bump you play some regular games.
First assume you lose all the subsequent games. You lose points. Say your opponents was 11k. Had you not had the “undue win” and stayed at 11k, you would lose about 15 pts, but because you are now 9k, you lose 20.
Or if you win the subsequent games. had you been 11k you would win 15pts, but since you are now a 9k you only win about 10.
The system fixed itself. From my estimation it takes 5-6 games to revert to your true worthiness.
Yes, but in reverting to your true worthiness you boosted up the ratings of the weaker players that you beat, because they got rank from your own wrongly boosted up rank.
In this way, that solution causes everyone’s rank to gradually increase. It’s not the problem of the individual personal rank being wrong that is causing this disussion.
It is the problem of the overall rank-pool increasing.
Note that I have only stated the argument, just here, as I have heard it. I have not delved into the maths to confirm that it is true, but for what it’s worth, this is the same reason that new people are given 13k rank, not 25k. If new people got 25k rank, then every new person would cause the whole rank-pool to increase over time.
That’s call ripple effect. For each layer it ripples outward the effects diminishes exponentially. From my example instead of winning 15 my opponent would have won 20. That’s hardly anything at all. If a 30 points distortion is reverted in about 6 games then a 5 pt distortion would probably revert in 1 game.
I’m having trouble determining if you’re appreciating the point I’m making, because your reply still refers to the distortion of the rank of an individual.
Each individual player’s rank tends back to the previous value, but there is a net injection of rank into the pool.
Or so the qualitative argument goes: that extra rank that you injected into the pool supposedly doesn’t “disappear”, it gets spread amongst everyone, so overall our total rank increases even if no-one’s skill does.
Once again, I will note that I am parroting the argument that “the other side” is making, as best I understand it. I have not applied myself to the maths to see if quantitatively it hold up, but I am concerned to see the debators actually debating the key points, and this is the one that I think needs to be tackled.
If you talk about the effect of serial timeouts on the rank pool, then you are debating the right issue. If you can demonstrate that this is a non-issue in a mathematically convincing way, then you will win the point that you are tryingt o make.
If you continue to talk about the effect on a single person, it will frustrate everyone because that is not the key point.
The key reason, the only reason that I am aware of, why the serial timeouts are not counted is because if they were counted they would “inflate the rank pool”.
The only way this argument can be countered is by solid maths.
if you want mathematical proof I will get mathematical proof. Here is a link of an excel add-on
https://www.add-ins.com/free-products/chess-ranking-assistant.htm
that could be used to build a chess club rating system. I will make up one when I get back to the office tomorrow.
For the mean time, if you are worry about distortion to the “rank poll”, then I ask the following questions:
what if the player who time-out was kind enough to resign the games before he leaves?
what if someone decided to let his kid play with his account?
what if someone who is reading this thread is so fed up he decided to register a new account to deliberately lose to 17k and win 6k
what if …
the point is, distortion, individual or poll, is inherent everywhere. You can’t block all, so why bother only blocking one?
Any half decent rating system can handle the noise. that’s why we have statistics.
