No mating material shouldn't win

I don't need to stand anywhere to point out that your example is wrong.

White to move, Bc5, if black refuses to move without this rule he loses, with this rule he draws. In other words black will move so he achieves a draw at least, with the rule black won't lose on time but it was a clearly drawn position.

If it happens that white is the one that runs out of time then it his damned fault for playing slower than black, if someone finds mate in X but loses on time they should lose on time as long as the opposing side has sufficient mating material with or without what KC proposes. Your example is "white can draw thus he should be granted a draw rather than losing on time".

My example had to do with the slippery slope of allowing computer adjudication of specific positions reached after the time control to affect the result.

The problem precisely that your example does not show that.

If an adjudication process is implemented, it would ONLY run on games where there is a sole knight, or sole bishop remaining for player A, and player B times out but with significant material.

These games are a corner case anyway, so the adjudication process would only be implemented on rare occasion.

Checking if there is a checkmate in 1/checkmate on premove isn't exactly very expensive.

I cannot think of many positions where Knight could checkmate without opponent making a helpmove that gives checkmate in 1 for Knight - and unless the Knight can do so on the next move, the box has to be broken and then intentionally reestabilished, at which point it can be argued that the side with sufficient mating material is stalling for time, which is better donBe through a series of checks or move to queen anyway with piece+pawn at least. While there does exists a lot of (sometimes intentional) blunders in 0+1 games there is still a differene between blunder and helpmate, even if the latter may be called one of the former.

I can think of one example where a Knight can win without helpmate: http://en.lichess.org/training/5506
But that is solved by checking for checkmate on the next move.

As for single Bishop, while that could be blocked, any move that has the opposing player able to do so is not a mate but at best draw - which is what this discussion is arguing for - thus only unblocked cases are interesting. While less restrictive it again needs to be boxed in with (1 chesspiece) the help of a pawn or opposite-colored bishop. The latter is a case of "possible checkmate thanks to helpmate in 1", alreay talked about above. And the pawn case needs another piece to force or bait 'stronger' king into corner while they still will have at least another pawn or it is a stalemate; thus any cases where there is a single Bishop versus a king that isn't in the corner of it's color can be dismissed; even those that have it in the corner must have at least 1 more piece in additional to the pawn (& potential bishop helpmate).

Now, there is a question how far one can extend this 'insufficient' consideration. Perhaps a zugwang should be checked when time runs out, too. Beyond that, some might want to let the threefold repetition/50 move rule take care of it, some might want to let two knights to be a draw. But that is beyond the scope of this question; which was to not let lone Bishop or Knight win on time unless it is mate in 1. I think I can agree with that, as it is also closer to OTB rules of 'arbiter'.

Let the 'arbiter' be a stockfish that checks if there is forced checkmate sequence with that lone bishop or knight. This should be more in the spirit of chess*.

*Do note that some people here do enjoy playing time with chess rather than chess with time from my talks in the chat, which means that not everyone agrees.

@Clarkey: You'll be hard pressed to defend that on logical grounds. I see the proposed solution as simply replacing one arbitrary rule with another; worse, the proposed solution is more complex.

Why do you think that they're called "arbiters"? It's because it's their job to resolve the arbitrary.

It is possible to program in a system based upon what I call the "volatility" of the position. It could be employed to ask "with the possibility for inaccurate play, is it possible for player B (the side with significant material) to be mated in short-order?" I programmed a very similar system for creating the puzzles.

KC is correct, in normal circumstances in OTB you can elect an arbiter to adjudicate the position and determine if within reasonable play, player A (the side with only a knight or a bishop) can perform checkmate.

Maybe what we implement is arbitrary, but it's no less arbitrary than using a real arbiter. Sometimes a judgement call needs to be made, sometimes it's right, sometimes it's wrong; but we're proposing a system that is hopefully more right than wrong as it's using computer analysis.

Will I be able to yell 'No progress!' next time somebody makes me play out a KRkr endgame?

Or a drawn KPk endgame that I sacrificed to get?

This isn't going to extend to KRkr endgames - not yet.

There's 3 fold rep, 50 move rule, and mutual decision on a draw for that stuff. Automating the decision making progress for non-progress calls is far less trivial than determining if a single knight or a single bishop has adequate mating chances to be considered winnable.