The 50 move rule shouldn't exist!

I would guess that a 1000 move rule would already make some theoretically won 8 man endings unwinnable, so this would almost certainly not solve the problem.

Strictly speaking KNNK is only usually a draw. You probably wouldn't be laughing too long if you were Black here:

                                                         White to play.

1. Castling should reset the 50 move counter as should failing to make an available e.p. capture. This should be part of any k-move rule including the current 50-move rule - it was just sloppily constructed in the first place. I don't understand where double check comes into the picture.

2. This is in fact the purpose of the rule. If a player hits the 50 move rule 2 moves before mate in KNBK say then he's slipped at least 19 moves and usually considerably more, depending on the original position and the quality of the defence. The rule is there to ensure a player can't f*rt about too much. The problem is not the fact that there is such a rule, only that 50 is too small -  in many situations it can also penalise accurate play. There can't be a one size fits all short of effectively scrapping the rule. That's why I suggest a sliding scale.

3. I agree that making specific rules for specific endgames would be impracticable. You don't want to carry around a book of k-move tables just to have a game in your local. I suggest k-move rules where the k depends only on the number of men on the board and is practically sufficient for that number of men.

4. A margin of 50% is arguably OK. In long endgames, if you make mistakes without throwing it away, then you're more likely to draw by repetition than under a k-move rule with a 50% leeway.

6. This seems to be a good argument for a k-move rule. It's to stop players making too many random moves.

That is not how chess works! It doesn't matter how many good (or bad) moves you make throughout the game. The very endgame is what matters. If someone made 3 mistakes or took 55 moves, but still successfully mated the opponent, he won fair and square, checkmate, he lost, game over. Chess isn't about penalizing accurate play, it's about what happens at the end of the game. Inaccurate play can just make it harder to win, but that's all. The rule is not there penalize mistakes. It is there to stop useless games from going on, nothing more. If a player has mate, he won, period. HE WON! The other player is basically saying "ohh this was a good game, i've been sitting here for 3 hours, it took him 60 moves to mate so it's not really a win, i think it was even"   NO!!  HE LOST!!!

Black can only end up there with cooperation. There is no way to force a mate because knights cannot TRIANGULATE. 2 knights vs lone king is a DRAW.

The position shown is not a draw (try analysing it with Stock fish). It's also two knights vs lone king (I've counted them). Therefore 2 knights vs lone king is not necessarily a draw.

This is different from saying that two knights cannot force mate against a lone king (which is not what you said). This latter depends on the conventional meaning of a forced mate as needing at least 2 plies.

But my original comment would apply equally well to KBNK and other endings. Even if he did get a half point out of it, one cannot but feel sorry for Kempinski having to play out 50 moves while Epishin tries to work out how to mate in the game shown here: https://en.wikipedia.org/wiki/Bishop_and_knight_checkmate#Grandmasters_failing_to_mate

1000 moves would have been just too much.

Actually, that's exactly what he said. He said "there is no way to force mate because knights cannot triangulate". Everyone agrees two knights and a king are a draw, unless at the very end of the game there is a forcing situation of one or two moves or the opponent somehow co-operates.

The point is he said that after my diagram was posted not before. The fact is that "2 knights vs lone king is a draw" is, as I said,  strictly speaking false.  It's only usually a draw.

Btw I've reached the position below (which is not drawn) on a number of occasions practicing KNNKP. It is two knights vs lone king (again you can count them) and it's usually the end of quite a long forcing sequence.

 I have no way of determining whether everyone agrees two knights and a king are a draw, unless at the very end of the game there is a forcing situation of one or two moves or the opponent somehow co-operates, but I would agree that two knights and a king are a draw, unless at the very end of the game there is a forcing situation (not necessarily limited to one or two moves) or the opponent somehow cooperates. (This is actually true of rather a lot of endgames.)

My point was that strictly speaking, "two knights and a king are a draw" is not the same as "two knights and a king are a draw, unless at the very end of the game there is a forcing situation ... or the opponent somehow cooperates". The latter includes extra words which modify the sense.

But that is the way chess works (and has worked for a long time). I thought that was what you were complaining about.

In the game Kempinski vs Epishin shown here:  https://en.wikipedia.org/wiki/Bishop_and_knight_checkmate#Grandmasters_failing_to_mate

Epishin had a winning endgame right up to the time Kempinski could have claimed under the 50 move rule. Had he invoked the 50 move rule would you say that it was penalizing mistakes or stopping a useless game from going on? I think most people would say, "both".

That's not true.  The rules of the game don't have to include mates that require over 500 moves.  The rules were never set in stone.  They have evolved.  Some change for logical reasons.  As we discussed in the stalemate thread, the rules varied (draw, win for one side, win for the other, loss of turn, illegal move) until the rule was finally standardized in the 19th Century.  You and I agree that logic requires that stalemate should result in a draw.  It would be illogical for it to result in a win for either side.  However, some change for practical reasons.  The rule allowing the option to move a pawn two squares on its initial move was introduced in the 15th Century.  It wasn't illogical to restrict a pawn's first move to one square.  Neither was it illogical to allow the two-square option.  I think the reason for the change was to quicken the pace of the opening and make the game more interesting.  You argue that the game should continue as long as mate can be forced, no matter how many moves it takes.  Your argument is not illogical; it's just impractical.  Keep in mind that these rules apply to games played in sanctioned tournaments and matches.  Because of that, limitations on the duration of games are necessary.  The purpose of time controls and the 50-move draw rule is to limit the duration of games.  One may argue that more than 90 minutes should be allowed for the first 40 moves or that there should be no time limit at all.  Similarly, one may argue that more than 50 moves (without a pawn move or piece capture) should be allowed to force mate or that there should be no limit on the number of moves.  But if there were no time or move limitations, tournaments and matches could drag on for months, perhaps making them economically unfeasible.  Such limitations on the duration of games may be offensive to purists but they're necessary to make professional chess viable. Those rules do not apply to your casual games offline. You and your opponent are free to disregard them.

A couple of points on the previous two entries:

1. FIDE rules don't necessarily allow a draw claim for mates that take more than 50 moves. There are very many games on record that run to more than 50 moves and finish in mate.

2. FIDE laws already contain variations for tournament play, so the form (if any) of a k-move rule or k-move rules need not be the same for tournaments and basic play. I don't think the constraints of organising tournaments should be allowed to affect the basic rules.

I agree but I don't consider the 50-move draw rule or time controls among the basic rules of chess, which include how the pieces move and capture, checkmate, stalemate and pawn promotion.  Changes to the 50-move draw rule or rules regarding time controls don't change the basic character of the game.  If FIDE wishes to change either, it must consider the impact on sanctioned tournaments and matches. 

I said I don't consider the 50-move rule among the basic rules of chess.  I wasn't referring to the section of the FIDE handbook that covers it.  I was comparing it to the most basic rules, such as how the pieces move and capture, checkmate, stalemate, pawn promotion, etc.  Those rules have remained the same for ages and changing any would fundamentally change the game.  That's not true regarding the 50-move rule.  I think it's fine as it is but FIDE could make it 75 or 100 without fundamentally changing the game.

If they made it 100 then a lot of people would be able to win KNNKP endgames that they can't win at the moment. This is true now but wasn't in say 1930..

I think the actual size is pretty intrinsic to the game, which is to say I would agree with the FIDE classification of the rule. If they reduced it to a 20-move rule for instance a whole bunch of won KPK positions would turn into draws. If they made it a 5-move rule all hell would break loose; you wouldn't be able to rely on winning a KQK and that would probably be the least of your worries.

The trend is for people to be capable of playing endgames that need a greater than 50 k-move rule to complete. Troitzky initiated the trend, and EGTBs are instrumental in continuing it. While the latter don't come up with any reasonable method unless you have access to the tables (e.g. you're allowed to consult Shredder - in which case the method is to do as you're told) they do provide a perfectly accurate (though not necessarily correct) opponent to practice against. 

For this reason I think the 50 move rule is an anachronism and needs replacing (but not discarding). I think the 50 move rule was in effect for what I would call ages before FIDE tinkered with it. I think they also tinkered with it (albeit ineptly) for the right reasons, but I think they tinkered it back for the wrong ones.

I don't believe I've ever played even a casual game that included 50 consecutive moves without a pawn move or piece capture.  If FIDE changed the rule to 1000 moves, it wouldn't affect me.  So I don't have a dog in the fight.  I just understand the necessity for reasonable limits on the duration tournament and match games.  If the overwhelming majority of endgames can be won within 50 moves (without a pawn move or piece capture) and only a relatively few require more than that, then I think the 50-move rule is appropriate.  But if a significant minority of endgames require more than that, I would support increasing the number.  But increasing it to 500 moves or getting rid of the move limitation altogether is silly.  That would satisfy extremists like EndgameStudy but why should that matter?  I would bet the vast majority of players would not want support a radical change in that rule.  I'm certain that most players, amateur and professional, have no interest in playing 500 move games.  If the 50-move rule causes one to draw a game that could've been won in another 50 moves absent the rule, so what?!  EndgameStudy's argument that the rule isn't fair is ludicrous.  A rule in effect before the game starts and applies equally to both players is fair.  

I think it might actually be the case that the great majority of endgames need a k-move rule with k>50.

This on the basis of endgames with random material rather than endgames that occur in practice.

But there is a catch 22 in the last sentence. Many endgames probably don't occur in practice because people don't know how to play them, and if they did it might well be unlikely they could win within the 50 move rule.

When Troitzky analysed the White won KNNKP positions he could only find three recorded games where it had actually ocurred (and I think one of these was actually agreed a draw the move before the ending actually appeared). Just a few years ago there were two such endgames being played at adjacent tables in the same American tournament and I think this occurrence and Troitsky's analysis are directly linked. People didn't convert into KNNKP prior to Troitzky's analysis because they didn't understand it.

In due course you could expect the same for other awkward endings as a result of EGTBs becoming generally available. This could only enrich the game, but the prospect would definitely require a change to the 50 move rule. People are unlikely to make the effort to learn an endgame they can effectively never use.

There are probably still more friendly games and individuals playing computer programs than tournament games. If it's a friendly then very often the game will be terminated by, "time gentlemen please" rather than any k-move rule anyway, but I still think it's important to keep tournament considerations separate from the game chess. Tournament chess has acquired some of the aspects of big business over the decades with sponsors and so on and obviously tactical coups and brilliances = bums on seats and the shorter you can legislate the ending to be the quicker your turnaround. The endgame is necessarily sadly neglected to start with because you don't get to play it unless you survive the preceding phases, but it is traditionally part of chess - it shouldn't be stifled for the convenience of the tournament circuit.

The answer is easy enough. Separate k-move rules for the basic rules and competition rules as is already the case. At least then it's clear that the separation is for the convenience of the organisers, participants and sponsors of tournaments rather than any consideration of the game chess.

And I agree that increasing it to 500 moves or 1000 moves or discarding it would be silly. What I suggested was increasing it to enough moves to accommodate the endgame being played with a reasonable reserve. It would remain at 50 moves for example for bishop and knight or any four man ending. And obviously if you're playing a casual game and you don't fancy struggling for 500 moves you can always say, "Sod this for a game of sowjers, are you feeling thirsty?".  You don't then have a result, of course, but I agree with EndgameStudy; in that case you don't have a result.

I'm not an extremist. I'm just saying it's silly to enforce the rule in positions that will me mate next move