The Longest Possible Chess Game | Hacker News Reader

The Longest Possible Chess Game

chess.com

79 pointsmonort16 comments

The article itself doesn't arrive at the correct number of 5898 moves, but some of the commenters do. At my own page http://tromp.github.io/chess/longest.html you can play through a 3-moves-shorter game.

It seems like the alternating of irreversible moves makes the 5870 shorter than it could be. The pattern of (49.5 + b) + (49 + w) + (49.5 + w) + (49 + b) + … could be replaced with several moves in a row by black, then several in a row by white, allowing some of the 49s to be 49.5 instead. Is that where the discrepancy comes from?

Yes, that would explain most if not all of the 56-ply discrepancy. In the game on my page you can indeed see black making irreversible moves on move 50,100,150, ... 450 and white's first irreversible moves comes only at move 500.

A user who is banned asked in this thread why you can't have an unlimited-length game where kings just move back and forth. The answer is

https://en.wikipedia.org/wiki/Fifty-move_rule

https://en.wikipedia.org/wiki/Threefold_repetition

All that the 50-move and threefold repetition rules do is permit one player to claim a draw; the game may continue if both players wish to play on. (It has happened, particularly threefold repetition where the repetitions are spaced far apart and the players don't have much time.)

There are actually new FIDE rules as of 2014 or so, where after 75 moves, or a fivefold repetition, the game is immediately declared drawn without the need for a player to take action.

This is a good point in that it means players could have collaborated to deliberately achieve an arbitrarily long game (and indeed, in some of the long reported games it looks like nobody chose to claim a draw for a while). I guess such collaboration isn't sportsmanlike in a competitive setting, but it wouldn't have been precluded by the rules.

Thanks for that clarification.

At the highest levels it probably does imply collaboration, but where I play at (much) lower levels, I might very reasonably choose not to take a draw because I think my opponent is likely to make a mistake that will result in a win for me. My opponent may reasonably think the same about me.

No, if no one claims a draw, it can continue forever.

In perfect play either black or white resigns in their first move or agrees to a draw.

But in in-perfect play you could get a situation where it plays until heat death. You at the very least could wait for a bit flip on the memory of the opposing player that might let you win.

Of course two Kings can't checkmate each other. If the board bit flipped, well that would be cheating / not a proper game. So in this specific case it'd be agreeed to a draw without the 50 move/repetition rule by the most basic of programs since checkmate is not possible given any play. But legally they wouldn't have to, unlike the above scenario when it can make logical sense to continue.

I thought the 50 move rule has been lifted for a few cases where a position is known to require more moves to win.

However, http://www.fide.com/component/content/article/1-fide-news/10... doesn't mention that, and even has the in some sense stronger "a game is drawn if any series of at least 75 moves have been made by each player without the movement of any pawn and without any capture.", which doesn't require either player to claim the draw, and, thus, cannot be avoided.

So, even if the players wanted it, a game can't be longer than somewhere around 96*75 or 7200 moves.

The 50-move rule was extended for certain types of endgames, but then it was felt that keeping a list of exceptions like that was kind of silly, and they reverted back to the 50-move rule in all situations.

It was up to 75 for a while, and then someone worked out a winning sequence that required more than 75 moves, and there is every reason to believe longer sequences are possible - to work them out is a brute force problem (chess is not mathematically solved)

Note that by winning sequence I mean perfect play on both sides - there are lots of sequences to allow to allow the game to continue longer than it has to.

In fact, there are already solved positions that are known to require over 500 moves to mate:

https://en.wikipedia.org/wiki/Endgame_tablebase#Tables

It seems like different parts of the chess world are a little bit divided about the best way to integrate this knowledge into the game.

From 2007, for whatever that's worth. I only noticed because the sudden mention of a "PDA" made me think, "what year is it?!"

One thing that I wondered about when looking at the longest recorded professional games that occurred in real play was at what point they reached a position that is actually solved in a tablebase. In particular, chess endgames involving only 7 pieces were explicitly solved by computer search in 2012. (But I also wonder if bringing those solutions to bear in a real game would involve ignoring the 50-move rule, since tablebase solution for these positions may often require more than 50 moves to complete.)

On tablebase-informed endgame theory, some of the professionals in these ultra-long games should probably have resigned long before the actual end of the games, except that they can also assume that their opponents don't know the full solutions.

Professionals may try to get a draw by getting their opponent to make an error under time pressure, even if they know their position theoretically is a lost one.

That even happened back in the time when games got adjourned after x moves (sort of an example: https://www.chess.com/blog/NimzoRoy/a-famous-bishop-vs-knigh...), so should be more popular nowadays, if the losing party has more time on the clock, or thinks to know the particular endgame better.

>Otherwise, we have nothing further to discuss here regarding the longest possible game.

The 50 move rule is there to ensure that the players don't have to occupy every possible position twice before making progress, but it is not the only possible way to put an upper bound on the game's length. Since the position will eventually repeat, and the three repititions rule can be invoked.

Reply on news.ycombinator.com