Timeout rules

"=> So, no, I did not suggest to ignore, weaken, whatever, the current implementation of the insufficient material rule."

And no, I never said you were unfairly favoring one *current implementation* on this site of one rule over the other.

I did however say you are favoring the priority of having one rule supported over the other, which you just admitted to here:

"About the "Insufficient Material" animal:
I think the current Lichess engine already does a good job in detecting those, and that there's no urgent need for improvement. This is what I wanted to say with: "Ignore the exact FIDE definition of the immediate draw rule", because that's exactly what the engine currently does. (If Lichess really exactly obeyed the immediate draw rule, then example #37 would be an automatic draw, rather then waiting for the players to agree on this.)"

That is basically admitting to one rule being more important to completely support to you than the other.

You know that both implementations are incomplete and incorrect (automatic draw vs. timeout draws and losses); you're saying you would rather ignore the problem of the insufficient material rule and evaluate on timeout.

"If no rule conforming implementation is possible, then there should be at least a simple "always draw" or "always lose". The latter one is probably the better one;-)"

Exactly.

It does indeed seem to be a FIDE regulation that the game is a draw on timeout if the opponent could not have checkmated as of that position, but I still think that the player who timed out in a game where only his opponent was powerless to checkmate, is at least as much of a fuck-up in his own right for timing out where he could do more.

So, even though it can happen, it doesn't change that he was retarded.

Both rules for drawing, automatic and time-out, are virtually impossible to support perfectly in a way that doesn't eat up server bandwidth, so I'd rather just say fuck it and count time-out as a loss. Equally as much, I have no plans to suggest a revision to the draw rule of Insufficient Material.

this happens once every 10000 games or so; so nothing hugely to worry about. besides, that's only 10000 games or so at my level; at lower levels the probability is too small to occur.

now...if i gained all the rating back from those cheaters....

>>And no, I never said you were unfairly favoring one *current implementation* on this site of one rule over the other.<<
Okay, now I understand your point.
Yes, I did favour a future implementation correctly supporting the timeout rule rather than the immediate draw rule.
But the only reason for this was that I saw no way of improving the current implementation of the immediate draw rule.

And I think you also admitted that the implementation of the immediate draw rule can't be improved:
>>The only solution I can think of would be to create a chess engine off of the Stockfish source which uses "Helpmate" mode to see if either side can commit suicide and be checkmated and externally request the answer from an engine after each player's move.<<
Such a check after each player's move would have properly triggered an automatic draw in example #37; but due to the limited server resources, it's not possible.

I favoured correctly implementing the timeout rule, because I thought doing such AI engine request after each timed out game were affordable.

But well, I changed my mind: My suggestion eats too much server resources as well, giving the large number of games which time out.

So we agree, a player who runs out of time should lose the game, no matter what the situation is on the board?

I went ahead and created a new thread in the Lichess Feedback forum, crying for Thibault to implement this:

lichess.org/forum/lichess-feedback/if-a-player-runs-out-of-time-they-should-lose-the-game-always

Xiaoqiao:

There are indeed cases were a game was treated a draw although it was lost by the timed out player (see post #1).
I'm not sure how the Lichess database of past games is designed, but maybe rectification of your games with "those cheaters" is possible....

unfortunately for me, the rating can't be changed since if it will, the previous rating of my game will be changed and my opponent will have faced a different rated player etc etc.

there are cases indeed, but i think they are way too rare to be an issue.

I think the implementation as it is currently is perfectly fine. I am not sure why people are beating a dead horse. If you don't have enough material, you draw. Our of those draws, probably about 1 in about 1 billion games is going to have a situation like that which originally created this debate, in which there is a potential victory by suicidal play from the opponent. So 1 in about 1 billion times there will be an error in resolution. That's pretty darn good.

In any other instance, i.e. there is sufficient material on the table to result in a win or loss under normal circumstances, the clocked-out player loses automatically. I'm not sure why there needs to be a change, as the most accurate manner of dealing with it is already implemented.

The problem with the current implementation is that it is intransparent, meaning that it doesn't correlate to a written version of chess rules.

For me it is not obvious why the site's mechanisms rated the game in post #1 as draw while rating games like in post #35 like a loss. So if one doesn't know the implementation details of this site, one can't understand why the FIDE rule is broken in both directions (draw instead of loss and loss instead of draw).

Codification of the rules is essential for fair play - I wouldn't want to play without that.
This doesn't mean that I insist the FIDE rules to be implemented on this site exactly, but at least the rules should be easy to understand, simple and straightforward.

"Timeout always means loss." is an example for an easy to understand rule.
The current implementation of this site, rating some games as draw (post #1) and some games as loss (post #35), both not conforming to any written version of the rules, is not an example of an easy to understand rule; it appears confusing and obscure to me. Thus it prevents fair play.

But the codification IS simple. IF you have enough material to win (the position is IRRELEVANT as in post #35) and your opponent times you, YOU WIN! If you DON'T have enough material to mate under NORMAL CIRCUMSTANCES (i.e. your opponent not being suicidal as in post #1) YOU DRAW! It's quite simple, it's systematic, and it makes perfect sense. I'm not sure why you want to drastically alter the implementation so that all time outs are losses, because all that does is further encourage someone who is in a lost position to try to trade down to their king and move around quickly and hope for a timeout when the best they should get is a draw.

In #35, white should have been a second faster and been able to snatch the rook up and get the draw. Simple as that. And quite honestly, there's no logical reason for that position to ever occur anyway. You are using ridiculously impossible positions under NORMAL play to justify your stance that implementation isn't transparent enough and needs changed.

It's quite transparent and makes perfect sense, and I am pretty sure it's fairly standard implementation for online play. If it is not, then I don't know why the "always lose" system is perceived as "better" because this is actually slightly CLOSER to FIDE's actual rules, barring the exceptional cases which would require arbitration in tournament play anyway!

Go troll somewhere else, the system's fine.

line 2: "and your opponent times you," should read "and your opponent times out," *

Also I find it absurd that there's been 5 pages of debate over a handful of odd situations so exceptional that one might occur once every year of five out of every game played on this site. It was explained, the explanation made sense, and most everyone stopped talking about it until you dragged it back up a week later and won't let it die.

Trying to seek some sort of glory for being "the guy who got rules implantation on Lichess changed because of a one in a million game scenario"? I think it's an ignorant and petty argument quite honestly, and you should just well enough alone. It isn't going to end the world when it gets asked again in a year or two when it happens the next time. :P

>>Trying to seek some sort of glory<<
damn! you revealed my plan! I wanted to apply for 2013's pettiest argument of the year award...
:-P

ok, now seriously:
>>But the codification IS simple.<<
The explanation you kindly provided in post #47 is indeed simple, but it still lacks a codification. "codification" means to write down the rules in some appropriate place.

>>If it is not, then I don't know why the "always lose" system is perceived as "better" because this is actually slightly CLOSER to FIDE's actual rules,<<
I considered the "always lose" system being close to the FIDE rules, as it just requires omission of a "however" clause. But whether one implementation is closer to the rules than another is arguable anyway.