Mate Threat Extensions

From Chessprogramming wiki
Jump to: navigation, search

Home * Search * Selectivity * Extensions * Mate Threat Extensions

Mate Threat Extensions are extensions caused by a mate threat of the opponent detected after making a null move in a fail-soft framework, or alternatively with a low shifted, or full window. Some programs detect mate in one threats statically, i.e. as part or byproduct of king safety in evaluation, to extend near or at the horizon.

Threat Extensions

Mate threat extensions are special cases of the more general Threat extensions not only restricted to mate scores but losing some material. According to Ernst A. Heinz, the Deep Thought team already used fail-low null move scores to detect threats before 1990 [1], descibed by Thomas Anantharaman in his thesis [2] and ICCA Journal paper [3], where he mentioned fall 1986 implementation in ChipTest.

Deep-Search Extensions

The term Deep-search extension was coined by Chrilly Donninger in his Null Move and Deep Search paper [4]. As pointed out by Ernst A. Heinz, Donninger's idea was to extend the search by one ply if a null move near the horizon does not fail high and the null move score plus a constant margin is below alpha, while the static evaluation indicates a fail high. He further states [5]:

Donninger's idea is to extend the search one ply if a null move near the horizon (e.g. at depths <= 3) does not fail high and the null move score plus a constant margin (e.g. minor piece value) is <= alpha while the static evaluation at the node is >= beta (i.e. fails high). In order to get meaningful results for the null move score, you need to do it with a full alpha-beta window instead of a zero window (this is a known error in Donninger's original article). 

DarkThought

Citing How DarkThought Plays Chess [6]:

In order to avoid possibly explosive growth of the search tree as caused by excessive deep search extensions in the case of repeated mutual mating threats, DarkThought restricts them to null moves at depth = 2 in the first "2 * iteration-number" plies on all paths. 

See also

Publications

Forum Posts

Re: mate threat extension/null move by Don Beal, CCC, October 04, 2004 » WAC booster

External Links

References

  1. Re: Deep Search Extension by Ernst A. Heinz, CCC, January 19, 1998
  2. Thomas Anantharaman (1990). A Statistical Study of Selective Min-Max Search in Computer Chess. Doctoral Thesis. UMI Order Number: AAI9111467, Carnegie Mellon University
  3. Thomas Anantharaman (1991). Extension Heuristics. ICCA Journal, Vol. 14, No. 2, pp. 47-65
  4. Chrilly Donninger (1993). Null Move and Deep Search: Selective-Search Heuristics for Obtuse Chess Programs. ICCA Journal, Vol. 16, No. 3, pp. 137-143
  5. Re: Deep Search Extension by Ernst A. Heinz, CCC, January 22, 1998
  6. Extension Heuristics from Ernst A. Heinz (1997). How DarkThought Plays Chess. ICCA Journal, Vol. 20, No. 3, pp. 166-176

Up one Level