It seems that one just happened in a game between Pascal Huybers and Thnikkaman. http://littlegolem.net/jsp/game/game.jsp?gid=1456870&nmove=169

I should say that because it was my lunch break, I was not able to play through the entire game, but I looked at several points, and it looked natural.

Yes, it's one to save for posterity all right. We're going to save it at mindsports.nl in any case. I've asked Pascal to agree with Thnikkaman on a final move (humanwise). Nice to see an odd result in the standings :)

It's been a long time since I played Havannah here, so I forget- are players able to propose and accept a draw? It seems more than a little inconvenient that you would have to fill the board, especially since the outcome was clear many moves earlier.

It would seem that the eventuality wasn't foreseen in Havannah.

Is this the first known occurrence?

As far as I know there were a couple of bot-draws on very small boards.

[game;id:1456870;move:169]

You might want to post a position around move 75 or thereabouts, when it became clear the game was a draw. The 'filling in' stage was only required to force the system to a result. An option to offer a draw would have made that unnecessary.

[game;id:1456870;move:75;title:As of move 75]

gg, it was a fun game. The end was a bit boring though. I actual wanted to resign when I saw that I couldn't win. But then I realised that he couldn't do it either.

That's a remarkable game. I would've been 99.9% certain that a draw was impossible on a base-8 board. Christian is right, too: the only draws I've ever noticed on this site have all been on very small boards, and when both players were bots:

http://www.littlegolem.net/jsp/game/game.jsp?gid=1153952
http://www.littlegolem.net/jsp/game/game.jsp?gid=1299376
http://www.littlegolem.net/jsp/game/game.jsp?gid=1290125
http://www.littlegolem.net/jsp/game/game.jsp?gid=1260811

It occurs to me that it's worth having a better player than me ask what the last move that isn't definitely best was. For instance, 72-75 are part of the same drawing sequence. Each move is forced. So while the draw becomes obvious at 75, it is assured before then.

How early can we see a draw in this particular game?

I would also e interested in if Pascal or I “misplayed” and could have forced the win at some point. Wouldn't it be interesting if Havannah drawed when played optimally? Surprising but interesting :)

On size 4 there is a forced draw after 4 moves or so, but not from an empty board. There are certainly winning and losing opening moves on bigger boards, but it's possible there are also some moves that lead to a draw. It's actually quite likely that both of you were in winning positions at some point in the game, though it'd be so many moves from the end that it's unlikely that we'll ever have the cpu power to prove it.

hehe yeah but maybe someone had an intuition and could say like on move 20 P1 had an advantage but he thrw it away when he on move 49 allowed P2 to do etc..

It wouldn't necessarily be surprising if optimal play was a draw. At least with the swap rule, it is surprisingly plausible, since if there is any move from which you can force a draw, that is the best move for the first player, even if draws are very rare in normal play.

More specifically, any first move from which the second player can force a draw but cannot force a win.