Talk:Null-move heuristic
From Wikipedia, the free encyclopedia
 |
This article is within the scope of WikiProject Chess, a collaborative effort to improve Wikipedia's coverage of chess. For more information, visit the project page, where you can join the project and/or contribute to the discussion. | |
| Start | This article has been rated as Start-Class on the quality scale. | |
| Low | This article has been rated as Low-Importance on the importance scale. |
This technique can in principle be used in any competetive game, but as far as i know it is only useful in chess. Does anyone know another game where it works well? If so, this article should be reworded more generally and moved to the 'Game artificial intelligence' category Bouke 12:47, 23 February 2007 (UTC)
Better explaining why the null-move heuristic does something and why it is useful seems like it would be worth the time. I am not an expert on programming or this particular topic (though I have an education in game theory and know C++ from once upon a time majoring in CS), but as I understand it, from a practical perspective, what happens is:
1.) Before we look at our position, we assume that we pass our turn and look at what the best minimax score of the best sequence of moves is from a 'passed' position, with less depth, which is assigned a minimax score.
2. All moves with a worse immediate position score than doing nothing, are pruned. This saves a lot of time.
Now what concerns me about my understanding here is that, if this is correct, a problem with zugzwang could be that it prunes all possible moves because they fail the test for being better than passing the move, which can be fixed by simple coding identifying said situations as zugzwang... but does not lead to any disaster after the error correction. In other words, if it looked far enough ahead to see all positions pruned by the null-move heurstic, you could add code to identify that as zugzwang and do a standard search, right?
It could immediately identify that with all positions pruned the position was zugzwang by definition, and see that ahead. But apparently the flaw of this program is that it causes blindness to zugzwang, which suggests something is wrong about my understanding here. Would be happy to expand the article if I could find some clarification.
71.145.159.2 07:31, 13 July 2007 (UTC)
