If a player runs out of time, they should lose the game, always!

Article 6.9 of the FIDE "Laws of Chess" ( http://www.fide.com/fide/handbook?id=124&view=article ) states:

Except where one of the Articles: 5.1.a, 5.1.b, 5.2.a, 5.2.b, 5.2.c applies, if a player does not complete the prescribed number of moves in the allotted time, the game is lost by the player. However, the game is drawn, if the position is such that the opponent cannot checkmate the player’s king by any possible series of legal moves.<<

A lengthy discussion ( http://lichess.org/forum/general-chess-discussion/timeout-rules ) showed that the last sentence of this article can't be implemented correctly.
So the suggestion is to completely ignore this last sentence - that is, running out of time always means losing the game, no matter what the situation on the board is.

This is simple and straightforward, as opposed to the current implementation which results in some games being marked as draw, although there's actually a loss (see

Thanks!

I once thought the reason it was easy to attain a high rating on this site was because there were not a lot of people. Not sure about that anymore.

gonna be difficult to find such a scenario.

Example setup:
White pieces: Kh8, Ng8, Pg5, Ph5
Black pieces: Kf7, Rc6, Nc5
White to move.

Assume that white is under time pressure, and blunders themselves into a forced checkmate sequence:
1.: Nh6+

Black starts the forced checkmate sequence against white:
1.: ..., Rxh6+
2.: xh6, Ne6

Normally white will be checkmated by this sequence, and every honest player would resign in white's position.

But what if white is an unfair player? White will do:

3.: (wait until the own time runs out)

The final position looks like this:
http://fr.lichess.org/y3zkowa8

So white gets away with a draw, although white deserves to lose this game and to be penalised for the unfair behaviour in the last move!

The problem with the current implementation is that it promises a draw on timeout; so people can abuse it and get a draw in a lost position. In my opinion that's not fair!

P.S.:
The forced mate sequence would continue:
3.: Kh7, Ng5+
4.: Kh8, Kf8
5.: h7, Nf7#
or:
3.: h7, Nf8
4.: h6, Ng6#

I think the problem with the time draw was practical rather than theory related. There are lots of ways that a checkmate could be forced with only a knight if the position were correct. EX: if stockfish engine sees a mate line for opponent, timeout is a loss rather than draw.

but this would be computationally expensive

I think Thib has discussed this before in another thread

CHESSCAPTCHA! :D

yeah, exactly
fide's damn timeout rules fail due to limited server resources.

but the status quo isn't fair either

Myself and Anon from #13 have agreed on a simple-enough all-around improvement to the rules.

I still think using my list for the draw on timeout (enemy has only a King or King + Rook vs. King + Bishop that can't checkmate) is equally as fair as the current state of the Insufficient Material draw rule which also doesn't cover everything, but I'll leave that preference up to him.

Fixing the system up by a bit is very easy to do.

Unless someone knows of another common type of position (which endgame pieces remain) that means exactly one side cannot checkmate but the other can, besides the two examples I gave.

And not worth doing. I've stated my opinion in the other thread as well. I think the system is fine as it is. It makes perfect logical sense to anyone with half a brain. You always lose on time, unless your opponent has insufficient material (i.e. very specifically defined as a king only, or a king and a single piece that is not a pawn, rook, or queen...the 2 knight mate is theoretically possible so it isn't considered insufficient material by FIDE standards) to secure a victory. In that case it is a forced draw.

There is only a single stipulation in the FIDE rules not taken into account, and that is position. If you can prove that while you don't have sufficient material, there is a smothered mate theoretically possible, you can have the draw overruled for a win on time. THIS STEP REQUIRES ARBITRATION, which we don't have here. So it's NULL.

i.e. the current system PERFECTLY reflects FIDE rules for rated and timed chess games which have no arbiter.

So......why do you want to change it again?

My interpretation of the above cited FIDE rule is that applying it to a game doesn't require an arbiter. It says "the opponent cannot checkmate the player’s king by any possible series of legal moves".
This can be determined by any chess player and doesn't require arbitration. Using the technologies of chess engines this can even be done automatically.

This FIDE rule is quite different from the corresponding USCF rule, the latter one explicitly requires an arbiter in such cases to decide whether checkmate can be forced. (Note: The FIDE rule just requires checkmate to be possible.)

I know, there were cases in a chess world championship where the FIDE rules also caused arbitration, and even an appeal. But as far as I know, this was only required because the opponent and the even arbiter didn't understand the rule themselves.

Wikipedia has a little information about that:
http://en.wikipedia.org/wiki/Monika_So%C4%87ko#Rules_appeal_in_2008

What? I said: "the even arbiter"? No, Sorry!
It should read: "even the arbiter"

Maybe I should better watch my word order...