Computer Chess Club Archives


Search

Terms

Messages

Subject: Definition of Depth, kns, total nodes per move, strength

Author: K. Burcham

Date: 19:53:28 05/12/02





I have read some posts here about running a tournament between programs using
search depth to limit each move. i ran these tests to show how i dont think
these results would be fair. with ponder off, and a limit of depth per move, set
to 12, it would be no contest between Deep Shredder and Deep Junior7.

I have always been curious about the total amount of data these programs cover
during analysis. and how these programs present this data to the user in kns,
depth, and total nodes per move in the engine window. Dan Corbit told me one
time, why did I worry about this---I dont, i am just curious about it.
Robert told me he didnt understand my question---i just didnt have my question
about this phrased properly. but you can see with this test why i am curious.
big differences in some info we get in the engine window. you can really tell a
difference when you run two pcs analyizing the same position, with two different
programs at the same time. all programs run to depth 21.

what else can we tell from this information?
kburcham




Depth 21 Test
Using the famous Deep Blue position,  37.Be4 in game 2.

 [D] r1r1q1k1/6p1/3b1p1p/1p1PpP2/1Pp5/2P4P/R1B2QP1/R5K1 w - - 0 1


2x1533 mhz

Deep Shredder Paderborn     152 minutes      551kns     5,035,356,609
Deep Junior 7                 4 minutes    1860 kns       447,474,000
Deep Fritz                   57 minutes    1420 kns     4,872,685,000
Fritz7                       51 minutes      938kns     2,870,359,000
Chess Tiger 14               23 minutes      497kns       686,862,000
Gambit Tiger 2.0             29 minutes      493kns       859,348,000


1000 mhz

Chess Tiger 14.0              28 minutes    306 kns          533,120,198
Fritz7                        82 minutes    583kns         2,870,359,000
Deep Fritz                   126 minutes    709kns         5,363,970,000



Deep Shredder Paderborn:    2x1533 mhz
432 hash
21.01	152:04 	+0.39 	1.Be4 Rcb8 2.Ra6 Qd8 3.Kh2 Rxa6 4.Rxa6 Rb7 5.Rc6 Kh7 6.Qa2
Qb8
7.Ra6 Qd8 8.Qa3 (5.035.356.609) 551.8

Deep Junior 7:      2x1533 mhz
432 hash
1.Be4 Rcb8 2.Kh1 Rxa2 3.Qxa2 Bc7 4.d6 Bxd6 5.Qa6
  ²  (0.58)   Depth: 21   00:04:29  447474kN

Deep Fritz:        2x1533 mhz
432 hash
1.Qb6 Rxa2 2.Rxa2 Ra8 3.Ra5 Rxa5 4.bxa5 Qh5 5.Qxd6 Qe2 6.Qb8+ Kh7
  ±  (0.94)   Depth: 21/51   00:57:46  4872685kN

Fritz 7:      2x1533 mhz
192 hash
1.Be4 Qd8 2.Ra6 Rxa6 3.Rxa6 Rb8 4.Kh2 Rb7 5.Qg3 Rb8 6.Ra7 Qf8 7.Qe3 Kh7 8.Kh1
  ±  (0.78)   Depth: 21/51   00:51:27  2870359kN

Chess Tiger 14.0:       2x1533 mhz
192 hash
1.Qb6 Rxa2 2.Rxa2 Ra8 3.Ra5 Rxa5 4.bxa5 Qh5 5.a6 Qg5 6.a7 Qc1+ 7.Bd1 Qxd1+ 8.Kh2
Qxd5 9.Qd8+ Kh7
  =  (0.11)   Depth: 21   00:23:05  686862kN

Gambit Tiger 2.0:        2x1533
192 hash
1.Qb6 Rxa2 2.Rxa2 Ra8 3.Ra5 Rxa5 4.bxa5 Qh5 5.Qxd6 Qe2 6.Qa3 Qxc2 7.a6 Qd1+
8.Kh2 Qxd5 9.a7 Qa8
  ²  (0.38)   Depth: 21   00:29:04  859348kN

>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
Chess Tiger 14.0:    1000 mhz
192 megs hash
00:28:58.7	0.11	21	533120198
Qb6 Rxa2 Rxa2 Ra8 Ra5 Rxa5 bxa5 Qh5 a6 Qg5 a7 Qc1+ Bd1 Qxd1+ Kh2 Qxd5 Qd8+ Kh7

Deep Fritz:       1000 mhz
184 hash
1.Qb6 Rxa2 2.Rxa2 Ra8 3.Ra5 Rxa5 4.bxa5 Qh5 5.Qxd6 Qe2 6.Qb8+ Kh7
  ±  (0.91)   Depth: 21/54   02:06:18       5,363,970,000

Fritz 7:    1000 mhz
184 hash
1.Be4 Qd8 2.Ra6 Rxa6 3.Rxa6 Rb8 4.Kh2 Rb7 5.Qg3 Rb8 6.Ra7 Qf8 7.Qe3 Kh7 8.Kh1
  ±  (0.78)   Depth: 21/51   01:22:14        2,870,359,000











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.