Author: Vincent Diepeveen
Date: 17:28:06 08/11/99
About theoretical branching factor.
Here an output of diep in a position with *a lot* of possibilities.
the number of nodes diep needs at small depths wondered me,
as i saw this in a blitz game this pos (moron-crafty or something)
and moron (diep at a celeron 550 dual NT) lost because it missed Ne5,
being in a big hurry.
However when i let DIEP search at analyzes level after some minutes
it's obvious that branching factor is *very* good.
Remember that i don't prune in DIEP except with R=3. I don't replace
nullmove last 4 ply by forward pruning either. Nothing of all that.
i don't do last ply pruning nor anything like that. Only nullmove R=3
in a recursive way using my double nullmove.
I do some extensions, although not too much. mainly check in this
version. My mating extensions are turned off for an internal reason i forgot.
The interesting thing is the branching factor. branching factor is
as we know calculated by just dividing times by each other.
So branching factor after ply=10 with the good score for Ne5 is excellent
for it here, though there are so many possibilities.
after some hours it gets again a little worse, i consider the problem here
that hashtable is overloaded. transposition gives so much, but with
an awfull overloaded hashtable it's good effect on branching factor is a
little less.
Looking forward to outputs of other progs here. Node comparisions
are cool too. The number of nodes diep needs here sucks bigtime
initially. Yet the branching factor it gets after that is unbelievable
well when using the blackman model.
I think it's possible to get this good branching factor because of 2
reasons:
a) transpositions
b) after 17 plies a lot of captures are done in that search, getting
to positions that have less possibilities, like endgames or even
pawn endgames.
c:\diep>diepm
Operator Time Set To 0 seconds
Book is on
Permanent brain is on
Allocated 39999896 bytes hashtables
# move freq games score*2 learn ==> huge cache
D I E P 1.60.05
black timeleft=27:46.40.00
r = b = r = k = ... 1 ...
o o = - = - = o ... 2 ...
- q - o - o - B ... 3 ...
= N = - b - = O ... 4 ...
- = O o - = - = ... 5 ...
= - = N = Q O - ... 6 ...
O O - = - = - = ... 7 ...
= - = - = R K - ... 8 ...
white timeleft=27:46.40.00
white to move
# move freq games score*2 learn ==> game cache
D I E P 1.60.05
clean
Clearing all Hashtables...
anal
Analysis mode is on
process 0: engineflags = 0 denktime=10000000 maxtime=10000000
00:00 73 (0) 1 -2.81 Nd3xe5 d6xe5
++ f3-d5
00:00 124 (0) 1 -1.12 Qf3-d5 Bc8-e6
00:00 133 (0) 1 -0.94 g3-g4
00:00 136 (0) 1 -0.32 Qf3-e4
00:00 149 (0) 1 -0.23 Kg1-g2
00:00 581 (0) 2 -1.18 Kg1-g2 Bc8-d7 Qf3-d5 Kg8-h8
++ f3-d5
00:00 1089 (0) 2 -1.12 Qf3-d5 Bc8-e6
00:00 3677 (0) 3 -2.25 Qf3-d5 Bc8-e6 Qd5-e4 Be6xc4
++ g1-g2
00:00 4807 (0) 3 -1.92 Kg1-g2 Qb6-c6 b2-b3 Qc6xf3 Rf1xf3
++ d3-e5
00:00 7056 (0) 3 -0.97 Nd3xe5 d6xe5 Qf3-a3
00:00 11671 (0) 4 -3.05 Nd3xe5 d6xe5 c4-c5 Qb6xb5 Qf3xf6
++ g1-g2
00:01 19274 (0) 4 -2.10 Kg1-g2 Qb6-c6 Qf3-d5 Qc6xd5 c4xd5
++ f3-d5
00:01 24966 (0) 4 -1.61 Qf3-d5 Bc8-e6 Qd5-f3 Be6xc4 Qf3-g4 Kg8-f7
++ f1-c1
00:04 64198 (0) 4 -1.56 Rf1-c1 Qb6-c6 Qf3-f1 Bc8-g4
00:05 88467 (0) 5 -1.50 Rf1-c1 a7-a6 Nd3xe5 d6xe5 Qf3-a3 a6xb5 Qa3xa8 b5xc4 Rc1x
c4
++ g1-h2
00:09 162117 (0) 5 -1.45 Kg1-h2 Bc8-d7 Nd3xe5 d6xe5 Qf3-a3 Bd7xb5 c4xb5
++ d3-b4
00:11 192666 (0) 5 -0.80 Nd3-b4 d4-d3 Rf1-f2 Bc8-e6 Qf3xd3 Kg8-h8
00:15 265567 (0) 6 -1.20 Nd3-b4 Bc8-h3 Nb4-d5 Qb6-c6 Rf1-c1 Kg8-h8
00:34 642824 (0) 7 -1.13 Nd3-b4 Kg8-h8 Nb4-d5 Qb6-c6 Qf3-d3 Bc8-h3 Rf1-f2
01:04 1194428 (0) 8 -1.73 Nd3-b4 Kg8-h8 Nb4-d5 Qb6-c6 Nd5xf6 Qc6xf3 Rf1xf3 Be5xf
6 Rf3xf6
++ d3-e5
01:30 1670366 (0) 8 -1.54 Nd3xe5 d6xe5 c4-c5 Qb6-c6 Qf3xc6 b7xc6 Nb5-c7 Bc8-h3 N
c7xe8 Bh3xf1 Ne8xf6 Kg8-f7 Nf6xh7
03:07 3429269 (0) 9 -1.54 Nd3xe5 d6xe5 c4-c5 Qb6-c6 Qf3xc6 b7xc6 Nb5-c7 Bc8-h3 N
c7xe8 Bh3xf1 Ne8xf6
06:18 6818386 (0) 10 -0.38 Nd3xe5 d6xe5 c4-c5 Qb6-c6 Qf3xc6 b7xc6 Nb5-c7 Bc8-g4
Rf1xf6 e5-e4 Nc7xa8
07:37 8359146 (0) 11 -0.38 Nd3xe5 d6xe5 c4-c5 Qb6-c6 Qf3xc6 b7xc6 Nb5-c7 Bc8-g4
Rf1xf6 e5-e4 Nc7xa8
10:34 11670757 (0) 12 -0.46 Nd3xe5 d6xe5 c4-c5 Qb6-c6 Qf3xc6 b7xc6 Nb5-c7 Bc8-e6
Rf1xf6 Be6-d5 Kg1-f2 e5-e4
18:53 21083408 (0) 13 -0.35 Nd3xe5 d6xe5 c4-c5 Qb6-c6 Qf3xc6 b7xc6 Nb5-c7 Bc8-b7
Nc7xa8 Bb7xa8 Rf1xf6 e5-e4
39:58 45729915 (0) 14 -0.05 Nd3xe5 d6xe5 Nb5-c7 Bc8-h3 Nc7-d5 Qb6-c6 Nd5xf6 Kg8-
h8 Nf6xe8 Qc6xf3 Rf1xf3 Ra8xe8
01:29:16 102213656 (0) 15 -0.30 Nd3xe5 d6xe5 c4-c5 Qb6-c6 Qf3xc6 b7xc6 Nb5-c7 Bc
8-e6 Rf1xf6 Be6xa2 Nc7xa8 Re8xa8
04:54:33 334025696 (0) 16 -0.05 Nd3xe5 d6xe5 Nb5-c7 Bc8-h3 Nc7-d5 Qb6-c6 Nd5xf6
Kg8-h8 Nf6xe8 Qc6xf3 Rf1xf3 Ra8xe8 Rf3-f7
07:45:37 531377444 (0) 17 -0.10 Nd3xe5 d6xe5 Nb5-c7 Bc8-h3 Nc7-d5 Qb6-c6 Nd5xf6
Kg8-h8 Nf6xe8 Qc6xf3 Rf1xf3 Ra8xe8 Rf3-f7
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.