Symbolic: BWTC.0001 (12 nodes)

Discussion of chess software programming and technical issues.

Moderators: hgm, Harvey Williamson, bob

Post Reply
User avatar
sje
Posts: 4675
Joined: Mon Mar 13, 2006 6:43 pm

Symbolic: BWTC.0001 (12 nodes)

Post by sje » Sat May 05, 2007 2:09 am

Symbolic: BWTC.0001 (12 nodes)

[D] r1b2rk1/pp1p1pp1/1b1p2B1/n1qQ2p1/8/5N2/P3RPPP/4R1K1 w - - 0 1

See: http://homepage.mac.com/chessnotation/BWTC0001.html

Uri Blass
Posts: 7993
Joined: Wed Mar 08, 2006 11:37 pm
Location: Tel-Aviv Israel

Re: Symbolic: BWTC.0001 (12 nodes)

Post by Uri Blass » Sat May 05, 2007 5:14 am

sje wrote:Symbolic: BWTC.0001 (12 nodes)

[D] r1b2rk1/pp1p1pp1/1b1p2B1/n1qQ2p1/8/5N2/P3RPPP/4R1K1 w - - 0 1

See: http://homepage.mac.com/chessnotation/BWTC0001.html
How do you choose your candidate moves?

I see that
Qxf7+ is one of the candidate moves but Bxf7+ is not one of them.
What is the reason?

I think that it is more logical to consider first Bxf7+ and not Qxf7+ that sacrifices the queen or Bh7+
Note that Bxf7+ also can force mate more slowly based on analysis

The only reason that I can understand for not prefering Bxf7+ in the search is the fact that Bxf7+give black 3 legal move when Qxf7+(like Bh7+ give black 2 legal moves) but a rule that always prefers searching first check that restricts the number of moves of the opponent regardless of material seems illogical and when it may be good for some tactical positions I cannot believe it can be good for tactics in general.

Uri

User avatar
sje
Posts: 4675
Joined: Mon Mar 13, 2006 6:43 pm

Re: Symbolic: BWTC.0001 (12 nodes)

Post by sje » Sat May 05, 2007 11:05 am

1. Bxf7+ was not included in the search because a forced mate was located first. The search concludes as soon as a forced result at the root is detected. The mate search does not attempt to find faster mates after a sure mate of any length is discovered.

No claim is made that the mate attack/defend metrics are good for anything other than mate attacks and defenses.

Uri Blass
Posts: 7993
Joined: Wed Mar 08, 2006 11:37 pm
Location: Tel-Aviv Israel

Re: Symbolic: BWTC.0001 (12 nodes)

Post by Uri Blass » Sat May 05, 2007 1:48 pm

sje wrote:1. Bxf7+ was not included in the search because a forced mate was located first. The search concludes as soon as a forced result at the root is detected. The mate search does not attempt to find faster mates after a sure mate of any length is discovered.

No claim is made that the mate attack/defend metrics are good for anything other than mate attacks and defenses.
It means that Qxf7+ is ordered before Bxf7+.
It is counter intuitive and as a human Bxf7+ is the first move that I look at it.

If you use SEE to decide about order of moves then
Bxf7+ seems better than Qxf7+.


Uri

User avatar
sje
Posts: 4675
Joined: Mon Mar 13, 2006 6:43 pm

Re: Symbolic: BWTC.0001 (12 nodes)

Post by sje » Sat May 05, 2007 8:17 pm

Uri Blass wrote:It means that Qxf7+ is ordered before Bxf7+.
It is counter intuitive and as a human Bxf7+ is the first move that I look at it.

If you use SEE to decide about order of moves then
Bxf7+ seems better than Qxf7+.

Uri
1. Qxf7+ comes before 1. Bxf7+ because it does more damage to the black king's flight square set. There is no SEE used for mate attack selection.

User avatar
sje
Posts: 4675
Joined: Mon Mar 13, 2006 6:43 pm

Re: Symbolic: BWTC.0001 (12 nodes)

Post by sje » Sat May 05, 2007 10:39 pm

I've updated the file on the web site. The output is almost the same as the first version except that Symbolic now outputs the average branching factor. And that's 1.22 for the given search.

User avatar
sje
Posts: 4675
Joined: Mon Mar 13, 2006 6:43 pm

Re: Symbolic: BWTC.0001 (12 nodes)

Post by sje » Wed May 09, 2007 11:01 am

I've updated the file on the web site again. The visible change this time is that the node marks have been expanded a bit and are now output to HTML in ASCII order to help with debugging.

Straight from the source, here is the rather flexible list of node marks:

Code: Select all

;; Node marks
;;

(defconstant TheNodeMarks
  '(
    IsAlternate
    IsBTM
    IsBook
    IsCandidate
    IsCertain
    IsCheckResponse
    IsCheckmate
    IsDraw
    IsDrawByFiftyMoveRule
    IsDrawByInsufficientMaterial
    IsDrawByStalemate
    IsDrawByThreefoldRepetition
    IsExtension
    IsFM
    IsForced
    IsForcedDraw
    IsForcedLose
    IsForcedMate
    IsFullyExpanded
    IsInCheck
    IsInDoubleCheck
    IsNoMovesAvailable
    IsOnlyOneMoveAvailable
    IsOriginal
    IsRoot
    IsSM
    IsSecondRepetition
    IsSingletonResponse
    IsTablebase
    IsTerminal
    IsWTM
  ))

Post Reply