Author: GuyHaworth
Date: 23:42:52 10/16/03
Go up one level in this thread
I surveyed the history of the 'k-move rule' in:
"Strategies for Constrained Optimisation", ICGA_J v23.1 pp9-20 (2000):
Further information is welcome if there are any omissions.
Ruy Lopez suggested a 50-move limit in Article 17 of his Chess Code of 1561,
perhaps in the interests of his fellow coffee-house professionals who played for
wagers. The 1883 London Tournament's rules, the basis of FIDE's rules today,
were the first to state that a P-push or capture would zero the move count.
In 1974, FIDE first enabled the 50-move rule to be varied.
They did so, with 100-move clauses:
- in 1978 for KNNKP
- in 1982 for KRP(a2)KbBP(a3) after the Timman-Velimirovic game
- in 1984 for KRBKR.
John Roycroft had argued for other endgames to be given more headroom but FIDE
did not agree with his suggestions.
By 1988, computer results were plentiful (Thompson), if not independently
2nd-sourced, and endgame-specific limits were suggested. However, FIDE adopted a
simpler stance, replacing the 100-move clauses by a 75-move clause for just six
endgames:
- KBBKN
- KBBKQ
- KNNKP
- KNNKQ
- KQP(x7)KQ
- KRBKR
KRPKBP with blocked Pawns ceased to be an exception at that time.
Stiller discovered KRBKNN's maxDTC of 223 in 1991 and so, in 1992, FIDE gave up
the chase and restored the universal 50-move limit.
KRNKNN's maxDTC = 243 was found in 1996 (and 2nd-sourced by Ken Thompson in
1999).
In the much more common situation with Pawns on the board, the move count can
still soar even with perfect play. KNNKP has maxDTZ = 82, a result in:
Tamplin & H, "Chess Endgames: Data and Strategy"
ACG-10 Conference (ICGA event, Graz, Nov. 2003) Proceedings
... and Marc Bourzutschky found, admittedly in a line minimaxing DTM(ate), a
move-count going to 165 in KRRPKQ. The actual maxDTZ of KRRPKQ is likely to be
of that order.
The Tamplin/H paper also surveys the DTC and DTZ50 stats, and gives examples
where minimising moves to mate, conversion or P-push is exactly the wrong thing
to do to preserve a win under the 50-move rule.
One could argue that chess-engine authors have every opportunity to pick up the
available EN DTM EGTs, or generate their own to DTM or another metric, and that
therefore the 50-move rule need not apply in C-C tournaments such as the
forthcoming ICGA WCCC event in Graz, Nov. 2003.
However, if the rule is to be dispensed with, competing authors should be given
reasonable notice of this.
I guess KNNKP and KQPKQ are the most likely to occur over the board, and of
these, KNNKP play is conspicuously the most affected by the 50-move rule.
Again, the Tamplin/H paper gives the stats.
It is unlikely that the maxDTZ = 243 of KRNKNN will be surpassed until 7-man
EGTs are computed. Establishing maxDTZ for 6-man P-endgames is however, a
target.
g
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.