Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: u2600 position analysis

Author: Will Singleton

Date: 12:23:46 02/24/99

Go up one level in this thread


On February 23, 1999 at 17:44:13, jonathan Baxter wrote:

>Here is TDChess' analysis:
>
>f7d7 3.27
>W: 3.27 d=2 n=227 s=17% n/q/e/t=65/21/63/0% p=82% b=0 mo=81% mtd=2 t=0.1 f7d7
>nps=2850
>c2c3 1.10
>W: 1.69 d=3 n=1115 s=12% n/q/e/t=65/15/46/2% p=92% b=0 mo=85% mtd=5 t=0.1 c2c3
>nps=12216
>h4f5 2.30
>W: 3.21 d=4 n=4756 s=8% n/q/e/t=65/16/29/4% p=95% b=0 mo=87% mtd=11 t=0.1 h4f5
>nps=39373
>f7d7 2.71
>W: 3.12 d=5 n=16725 s=5% n/q/e/t=74/13/37/12% p=95% b=0 mo=86% mtd=4 t=0.3 f7d7
>nps=71449
>W: 3.43 d=6 n=60934 s=4% n/q/e/t=71/16/36/6% p=96% b=0 mo=86% mtd=9 t=0.7 f7d7
>nps=105484
>h4f5 5.09
>W: 5.09 d=7 n=362279 s=3% n/q/e/t=62/22/48/4% p=96% b=0 mo=82% mtd=9 t=3.8 h4f5
>nps=117881
>W: 5.09 d=8 n=503981 s=3% n/q/e/t=68/19/43/6% p=97% b=0 mo=88% mtd=2 t=5.0 h4f5
>nps=121209
>f7d7 4.59
>W: 4.84 d=9 n=1805900 s=3% n/q/e/t=67/19/40/8% p=96% b=0 mo=82% mtd=6 t=17.7
>f7d7 nps=122162
>W: 5.00 d=10 n=7114735 s=4% n/q/e/t=68/19/40/8% p=97% b=0 mo=84% mtd=7 t=68.7
>f7d7 nps=123391
>
>The fields are: side to move, score (in pawns), depth, total nodes searched,
>percentage of positions that had a "stage change" in the eval, n/q/e/t are all
>percentages: n= successful cutoffs out of all cutoffs, q=quiescence nodes,
>e=extended nodes, t=truncated nodes; p=pawn hash table hits, b= number of brain
>entries encountered in the search tree, mo= move ordering, mtd= number of calls
>of the mtd loop before convergence, t=time in seconds.
>
>

If I read this right, it takes you 9 ply to see this, which is the same as
several other programs.  I wanted to try and get my program to see this earlier,
as some of the stronger programs do.  So I fooled with it and now it finds it in
8 ply, but I don't know if it's worth it.  It might play worse in other
positions, or slow it down generally.

In any case, here's the new 8 ply version:

Amateur 0.92 -- 7k/2pn1Q2/2r4p/p1qpp3/Pp1b3N/1P1P2PP/2P1R2B/7K w - -

ply   score     time     nodes      n/s    pv

  2     435     0.01       819    49140
  3     475     0.03      1735    34680  Qxd7 Qf8 Nf5
  4     477     0.09      5721    38100  Qxd7 Qf8 Nf5 Rc5
  5     475     0.26     19222    44340  Qxd7 Qf8 Bg1 Rd6 Qxc7
  6     502     1.49     96449    53040  Qxd7 Rf6 Qc8 Kh7
  6     503     2.50    151057    53280  Nf5 Qf8 Qxd7
  7     506     4.50    266634    55140  Nf5 Qf8 Qxd7 Rg6 g4 c6 Bxe5 Bxe5 Rxe5
  8     173     9.02    518960    57420  Nf5 Qf8 Qxd7 Rd6 Qxc7 Qxf5 Qb8 Kh7 Rg2
  8     522    26.09   1498077    57240  Qxd7 Qf8 Re1 Rd6 Qxc7 Bb6 Qb7 Bd4

Oddly enogh, the PV's are much shorter now, don't have the slightest idea why.

Will




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.