The 50 move rule shouldn't exist!

How many moves can a knight+bishop vs king endgame go on without any repetition, excluding triangulated positions?

A repeated position, but with the other player's turn to move, which doesn't count as repetition.

Ah. In that case obviously less than the million I so flippantly said, but obviously a lot more than it would be pleasant to play.

BTW, 2 knights vs lone king would be declared a draw immediately in a tournament, unless there was an obvious mate next move.

I think that's a local variant in your country, but I think the people I know might even agree a draw in that case (except for Stockfish of course).

There's no way to force mate with 2 knights vs LONE king unless it's mate on the move; and I don't know how the 50 move rule would affect the 2 knights vs Pawn endgame because there's a pawn involved.

It's quite true that there is no way to force mate except on the move from any KNNK position (that isn't already mate) but a mate can always be achieved by a series of legal moves, so under international laws the game would not be drawn unless the players agree a draw or one or other player were to claim a draw by repetition or (in competition play) one or other player after 50 moves were to claim a draw under the 50 move rule or the flag of the player with the two knights were to fall. In non competition play, since 1st. of July this year, if the players do not agree a draw and neither player claims a draw by repetition play continues until one of them dies. In the last case no result is specified.

The 50 move rule renders mate in some KNNKP positions impossible against correct defence if the defender (which I think may be either side) were to claim the draw, although it would be otherwise possible from the same positions. Mate may of course still be possible from these positions against inaccurate defence.

What's the maximum number of moves 2 knights vs pawn takes?

If it's not competition play (since 1st july this year) the the longest wins against accurate defence are 115 moves with accurate play. These are mates with the two knights and play is entirely within KNNKP. The longest corresponding wins for the side with the pawn are 74 moves but most of the play in these cases is within KNNKQ.

For competition play the longest mates (at least for the side with the two knights) are higher - I think 128 moves but my memory may be faulty; I can't play the ending taking the 50 move rule into account anyway.

 Wait, but the pawn moves at some point, so that would reset the move count, right? Then the rule (I still think time limit is better), should be 200 moves (400 plys) for the 2 knights vs pawn endings.

If Black wins from a KNNKP position then the pawn always moves unless that position is already mate. If White wins against best defence then the pawn usually moves, but not always. E.g. White can win from the position shown below, but cannot afford to let the pawn move.

                                                     White to play.

In the majority of drawn positions Black need not move the pawn

The pawn moves do quite often mean Black has no opportunity to claim a draw under the 50 move rule during long White mating sequences. For example I think all the White wins with a DTM in the range 100-115 can be played by White without incurring a draw under the 50 move rule, but that is only from practice against Wilhem/Nalimov; DTM. Conceivably the Sysygy DTZ50 database would manage a draw under the 50 move rule for some of these endings. Similarly some of the longer Black wins may avoid incurring a claim owing to the pawn moves and possible captures in KNNKQ, but I can't reliably comment because I'm too flaky on the Black winning play in KNNKQ.

The current claimable number of moves in competition play is, according to FIDE, 50 moves with a mandatory draw after 75. The 50 move rule I can say is definitely too low for White to win some positions that are winnable if the rule is waived. I think the 75  move rule is also too low; I would say somewhere around 85 moves would be required against (traditionally) perfect play for positions where White cannot prevent a Bishop's or central pawn from reaching its 5th. rank. I can't state an exact figure.

I don't know how you arrive at 200 moves. If Wikipedia is to be believed (which may be ill advised) FIDE at one time had a 230 move rule (2x115) in effect for the ending. I'd say if the rule is not to be completely altered it should be at least 110 (85+25%) which would bring it in line with the leeway allowed in KRKN, which I think is probably the longest 4 man ending. 

I take your point that the purpose of the 50 move rule is not to prevent a player from winning a won position that he's worked to acheive. I even take your point that the purpose is not to terminate play in positions where progress is being made even if that progress is less than optimal. I just think your whole game time limit is a daft idea that would acheive exactly the opposite of what you want. 

I was saying the time limit would be better because since the moves are so variable for these endgames, that focusing on the actual issue, which is time (everyone waiting for the next round, breaks between rounds..etc) would solve that problem. When people are waiting for the next pairings, they don't care how many moves are going on, just how long it's taking. Players would also be able to optimize their play by speeding up their play in easy positions, and slowing down in complicated positions. A move limit is saying even if x moves were played in 20 minutes and there's still 3 hours before the next round, the game has to end immediately anyway, which is not necessary. Besides if the next round HAS to start at a certain time, they might have to stop the game even BEFORE 50 moves anyway. The time limit would solve the waiting problem, but allow many more moves when there is plenty of time before the next round or the end of the tournament.

The issue of the 50 move rule is not and never has been anything to do with time. The rule is designed to terminate games which are making no progress but one player is refusing to agree a draw. That's what the caveats about pawns being moved and pieces being taken are about. It is assumed that if none of those things are happening for fifty moves nothing useful is going on. How long the moves take is not a part of the rule.

Your whole game time limit effectively puts a limit on the number of sensible moves that can be made in a game. Such a limit has never been in the rules. The effect would be only that in the longer more closely fought games the big time limit foot would come down; squelch, "sorry folks you're drawn". This irrespective of whether one of the players has built up a game winning advantage. The rule would rob dozens and dozens and dozens of times more players of well deserved wins than the 50 move rule ever could. The 50 move rule is hardly ever encountered in games. 

Is there not a "50 post" rule to these blogs 😂😂😂

Some of these blogs can take 534 posts! But that's OK we just speed up our posting rate a little and make a post every two minutes.

In positions like this:

U know there will never be progress here, BUT

You don't know if progress is being made because it is so subtle. Gradually trapping the knight and then taking it, winning with 2 bishops. If the time rule allows well over 50 moves to be played, how would it be robbing more people of wins, if it's allowing more moves? There's also the standard time control, so it's not like someone could spend an hour on 1 move in either case.

How could u have a 50 post rule when there are millions of people on chess.com LOL. There's one thread that has over 22,000 posts.

Obviously only in blogs where an argument is clearly drawn 😉