2011 Arimaa Man vs. Machine Match

The 8th Annual Arimaa Challenge Match is now in progress. The games will feature live commentary by high rated Arimaa players. Check this page for your local game times:

Local Game Times

The challenge prize is currently $11,000.

3rd game of the Arimaa Man vs. Machine challenge match in progress here

5th game of the challenge match in progress with live commentary here

In the 5th game not even a single piece was captued by the bot! Very nice to have a game that can't be simplified easily.. but what about draws? :)

Afaik they almost never happen, because if you make a move that repeats a position the 3rd time you lose. It is also forbidden to not change the position. So if a game is played with unlimited moves there will always be a winner. In Arimaa the maximum number of moves before the game ends in a draw is a setting like the move and game time that's agreed upon before the game starts.

Well, i know rules.. and, btw, one of these two games ends because the bot can't repeat the position, it seems to me.. but my question can be stated more formally in this way.. why most games end in a reasonable number moves? why rabbits never end face to face in the middle? :)

I have two questions about this game. One, how do you pronounce "Arimaa"

Two, can a piece move through a trap in two steps? I know you can do this if you have a friendly piece next to the trap, but can you move through a trap if you do not have a friendly piece next to the trap?

No you can't because traps are active even between every step. If I remember well, every rule works this way except the goal and the no-repetitions rule. On the other hand, sacrificing your pieces (including elephant) is sometimes used to clear the path for a rabbit! :)

If you want to learn the rules or for the pronounciation look this short video

Other languages are also available at arimaa.com

Btw the game is called like this because Omar Syed created it together with his son Aamir, therefore there are the 2 'A's at the end.

Rabbits never end up face to face in the middle because this is 'bad' play. It turned out to be better to forward the rabbits later in the game (otherwise they can easily be captured). Especially the rabbits can't move back rule makes that the game somehow automatically progresses. The only reasonable draw I see is when both players just move their pieces around without forwarding (or pulling) rabbits and without making captures.

I agree the problem I'm complaining about is extremely unlikely to arise out of normal play. I think that a rule like "After 1000 half-moves the game is a draw or silver win" would be clearer than the time limit and the odd scoring producedure contained in the "Math rules". By the way, has the scoring procedure ever been used to decide a game?

Not "Math rules" but "Match rules". They're online at: http://arimaa.com/arimaa/learn/matchRules.html. Best regards.

And not "After 1000 half-moves the game is a draw or silver win" but "After 1000 half-moves without a rabbit move the game is a draw or silver win". I mean, same problem as in Chess, same solution?! :)

lazy, thanks for the clarification on the rules. 1000 half moves seems extreme. Maybe 100 half moves without a rabbit move, same as chess. That should be sufficient, don't you think?

Ray, yes with the rabbit move rule maybe a more manageable number like 100 is preferable. But it is a really tiny detail if it never happens in normal games.

> but what about draws? :)

Draws used to be possible in Arimaa if both players lost all their rabbits, but it never occurred in any well played game. This was quite ideal because I had always felt that a "well designed" game should give an equal chance to both players such that a perfectly played game results in a draw, but the margin of draw should be very low so that even high quality, non-perfect games do not end in a draw.

Then one day Fritz Juhnke (author of the "Beginning Arimaa" book) pointed out to me that just because a game allows for drawn positions, it doesn't necessarily mean that perfect play will result in a draw. In fact it is even possible that high quality games result in draws, but perfect play leads to a win for say the first player. An example of such a game is Connect 4 where well played games result in draws, but since Connect 4 has been solved we know that perfect play leads to a win for the first player. This means that the game designer isn't off the hook just by allowing drawn positions; the game still has to be solved to say for sure if perfect play will lead to a draw or not.

Even though ideally I wanted Arimaa to be a drawn game if played perfectly, in practice I was finding that for elimination type tournaments (like the finals of our world championship tournament) I did not want there to be any possibility of draws, because another game would have to be played to break the draw and that makes the tournament schedule less predictable. So for tournaments we were using the rule that if a players loses all the rabbits, they lose the game. Eventually in July 2008 I made this part of the official rules and eliminated the possibility of draws.

Some people feel that games which do not allow for draws are flawed because we know for sure that one side or the other has won right from the start of the game. Fritz has convinced me that it doesn't matter so much if the game is ideally a draw or not, what really matters is that for non-perfect play the game should be balanced for both player. In other words for non-perfect, high quality play one side should not have a significant advantage over the other. Fortunately Arimaa seems to have this property.

Fritz Juhnke will be playing the 9th and final game of the challenge match tomorrow. He mentioned that he will be playing a cat handicap. This might be equal to playing with a knight handicap in chess. You can watch the game with live commentary from this page.

By the way, has the scoring procedure ever been used to decide a game?

During the Online Festival, the team games had a very short total-game-time clock, and I believe both games were won by score. Certainly the one I played in was.

Rozencrants, interseting :) So if time limit is short enough, then the relative value of pieces change! This is why i think that using match rules to ensure termination is odd! :)

Hmm, now that i think about it again, rabbits in Arimaa can't be used to ensure termination like pawns in chess because they can also be pushed backward. So i go back to my original proposal, to cap number of moves. Once the cap is reached, maybe scoring could be a good idea.. so that Gold can trade bit stronger pieces with bit weak pieces to advance the position..