Life In 19x19

I think this thread is getting out of hand. It was another rant against the escaper system at KGS, which wms has stated over and over that he doesn't want to change.

Any serious proposals should be started in another thread?

Anyone else, please calm down before posting.

xed_over wrote:KGS escaper system, as far as I understand, is by design. And so, by design, it is not flawed. It may still have its pros and cons, but its still by design.

Maybe that design is antiquated by now. People have stable internet connections nowadays, even in asia. Most unfinished games will be escapes. So maybe the design became flawed over time, the trade-off becoming bad.

If the escaper system was opt-out (i.e. per game setting for escape handling), escaping would die out very quickly, and the last four people on earth with a modem would lose 2 more games a year.

And regarding adjourning games that's a completely separate matter, which could remain regardless of escaper handling.

But we all know KGS is for kibitzing, so game results don't matter anyway.

xed_over wrote:
But the issue here is that many people assume the KGS escaper policy is flawed, while others do not. And this is where Redundant's reductio ad absurdum falls down, his argument assumes the policy is flawed. But if its not flawed, then it can't be compared in the same way, and we should learn to live with it, as Bantari suggests.

First, thank you for disagreeing with my by attacking my argument's premises and inferences.

I concede the point that my argument assumes the system is flawed. This is a debate in and of itself. However, I don't think Bantari was arguing that those who think the system is fine need to learn to live with it. He was responding to those who don't like the current system, by telling them to stop caring.


Neat side note: I actually think the KGS system is not deeply flawed, only somewhat sup-optimal. I dove into this mostly because I don't like the idea that one should give up on trying to fix a system they feel is broken.

PS: KITTEN

I have no opinion on which way to deal with the escaper problem, however it does seem there's a minor fix which would help with the specific situation SpongeBob had: don't let escapers start a new game for <insert some time period here> while their old game is still open.

Nice! We go from escapers to Rwanda to tortured Latin to religion to Robert Jasiek to Klingons and then to kittens.

Works for me.

All roads lead to Robert Jasiek. You should have seen that coming

I am with Bantari here: I don't let it bother myself.

Apart from that: Playing on IGS is fine, too. Why bother with KGS if you don't like the escaper rules.

Tryphon wrote:

xed_over wrote:

kirkmc wrote:You can't that's why I look for a trend. If there's only a couple, then it's clear it's the other players. But I've seen players who have lots, or who have forfeits in the game list, which show that they lost by forfeit because of escaping.

I know players who have lots of unfinished games in their list, but none of them were escaped by that player -- so you still can't tell.

Probabilities agree with kirkmc. And I do, with them. Too bad for the players you know.

That's not very good for me. I NEVER escape, but I have had as many as 7 or 8 italic games at a time, often within the same week. This is because when I get the time, I play a good number of games within a one week span, and the EGR seems to be filled with 4k-8k escapers.

This is the reason I am primarily playing Automatch games lately. I have found far fewer escapers doing that.

The escaper policy issue is an interesting issue, even if it is a bit over-discussed. As with many issues it seems to evoke a strong emotional response from many of the participants. While this is one of those issues where there is nothing new under the sun, I for some reason feel compelled to throw in my 2 cents (after all, it's a slow day at work...). Based on what's been said here and what I have seen stated other places, I may end up making assumptions and guesses as to the meaning of the posts of others, so I apologize in advance if I put words in anyone's mouth, or do not accurately describe someone else's intent.

When looking at how one wants an escaper system to operate, there obviously must be some core values and assumptions. I think it's differences in these values and assumptions that lead people to disagree on this subject*. For the KGS escaper system, I would reckon the main tenants are something like this (note: I number them for reference sake, not for any sort of priority)-

(1) The system should be completely objective and automated: There are a number of reasons for this. First of all, having an individual or individuals regularly changing the results of rated games allows too much room for mistake, abuse, and outside criticism, and this is an option best not explored. Also the sheer volume of work this task would require to implement, monitor, and quality control, puts out outside the range of practicality.

(2) The vast majority of all actions taken by players on the server are made in good faith: If you are to approach a situation on the server with no background knowledge of the people involved, all other things being equal you should assume whatever they have done their aim was to play an enjoyable game of go. I include this first of all because from my experience in KGS this is true; most people are there to have fun and play go. Further, this type of person would presumably be the target audience of a go server such as KGS (the aim of the server first and foremost being somewhere to play and watch go), so one would imagine designing the system for one's target audience.

(3) When possible, the implementation of the system should minimize negative impacts on other aspects of the server: The first thing that comes to mind here is the rating system. The goal of a rating system is to accurately estimate the probability of a given player winning should they play a game against another. When looking at potential systems you will want to choose one that will not negatively impact this estimate.

(4) The system should minimize the effort required on the part of the user to obtain the desired effect: The more work one must do in order to make a system operate, the less likely that system will operate efficiently or properly. Likewise, the more work the average user must do in order for a system to work, the more they are inconvenienced by the addition of that system.


When looking at these criteria, I would say the KGS system works quite well. The 1st is fairly easy to achieve and in fact most suggested alternate systems use it. The 4th is quite straightforward as well, and just about goes hand-in-hand with (1) as automated systems generally require little user input. I believe it's that 4th point that Bantari was referring to with the KGS system in that as an average user, in order for the system to work, the best thing to do is simply continue playing games and let it work. The system is designed to work without additional effort from the user, and operates quite smoothly when it is allowed to work things out on its own. There are a number of behavior patterns I have always found quite puzzling in watching how people deal with escapers, some of which are even listed in this thread, for example: Keeping a list of known escapers in your profile, looking at others profiles in advance to see if they are an escaper, and announcing to chat rooms that users are escapers. All of these behaviors work toward minimizing the number of games played with escapers, yet that is the exact condition under which the KGS system will not work. If all escapers play (and escape) a large number of games, the net result is that all escapers resign the games as soon as they leave (an outcome generally stated as desirable by critics of the KGS system), yet all of these behaviors listed work toward actively decreasing the likelihood of that outcome. If instead players put no effort into "combating" escapers, the system would work as designed and all would be well.


There is a common counterargument here stating that regular escapers will simply make new accounts, and continue escaping-- thus never resigning their escaped games. This however fits perfectly in line with points 2, 3, and 4. By constantly creating new accounts, the escapers will have difficulty creating a solid rank, therefore any effect their games may have possibly had on the ranking system will be minimized. Furthermore, requiring the constant generation of new accounts means that a much larger effort is required on those who wish to fall outside of KGS's target audience (those acting in good faith), while the effect on those who simply have connection problems (and are still acting in good faith) is minimized. Likewise, the tendency of this system is to exclude data points where the appropriate outcome was unclear, again reducing any bias introduced into the rating system.

I have a feeling it is points (2) & (3) are where most people differ in their opinions compared to KGS. At some point you have to just agree to disagree. Hm, that got a little longer winded than I expected, oh well.... Wms or Bantari if I've misstated or misinterpreted something either of you was saying (or what I've seen said in the past), feel free to correct me.

*I really think the main disagreement had to do more with emotional involvement, and the desire to have an extra win tallied up in a profile, rather than a rational disagreement over values.


[edited for grammar]

Mef wrote:...
When looking at these criteria, I would say the KGS system works quite well. ...

It's interesting that I guessed the conclusion of your post before I read it.

I'm just kidding, though. I agree that a lot of the debate may be due to emotional aspects.

When I look at my unfinished game list I see 6 of them. And some of them might even have been left by me due to wlan failure.
Afaik the list lasts 6 Months(~600 games for me), which means that in about 1% of my games the opponent escaped and did not get a forfeit yet.
So I don't think there is an escaper problem on kgs.

I like how wms liked the above post

Kirby wrote:I agree that a lot of the debate may be due to emotional aspects.

I fully agree. For me, this is purely an emotional issue. When I play Go, emotions are involved. When I win the game the game, this feels really good. If I loose the game, it hurts. (Some Go player had a T-shirt with 'If you can't stand the pain, don't play the game.' on it - have to get one for me )

And, guess what, if somebody escapes on me, I get angry. If he stays on the server and happily starts a new game while I wait for him to return, this makes me even more angry. He leaves the game to avoid the loss and the system allows him to do that. If the system wouldn't, him leaving the game would be essentially the same as a resignation.

There is a reason that the KGS system is so 'unique'.

SpongeBob wrote:And, guess what, if somebody escapes on me, I get angry. If he stays on the server and happily starts a new game while I wait for him to return, this makes me even more angry. He leaves the game to avoid the loss and the system allows him to do that. If the system wouldn't, him leaving the game would be essentially the same as a resignation.

There is a reason that the KGS system is so 'unique'.

the admins have actually told me that they think the player has the right to leave the game if they dont want to finish it. so theyre telling me i have the right to escape if i want to.

its part of the reason why i wont ever support KGS until something is done about the system. yes i have KGS+ but its only a consequence of being in the insei league. i think KGS is a good server but the admins do need to wake up and listen (no matter what they tell you, they dont)

xDragon wrote:i think KGS is a good server but the admins do need to wake up and listen (no matter what they tell you, they dont)

They listen. They just don't agree with you.