Author: Christophe Theron
Date: 09:22:37 02/08/04
Go up one level in this thread
On February 08, 2004 at 06:55:43, Eelco de Groot wrote:
>Hi Christophe,
>
>Success with Chess Tiger X! Does it still use the hybrid set-up, I mean with
>part of the engine not knowing all the rules of chess?
Yes, it is a faster engine.
>I know it is not the way you have been working on Tiger, but I still wonder if
>people would be interested in having someting like a personality editor for
>Tiger. I know I have had a lot of fun with that in Rebel but maybe that is
>personal. Still, I think it is worth considering for the future? Personally I
>don't find the SSDF numbers most important, less important than the quality and
>"feel" of the analysis lines, which is not exactly the same thing?
I consider that a personality editor would be a nice thing to sell more, but it
is a gadget. I focus on improving playing strength. A personality editor would
please people who would create a new personality, play a 20 games match against
whatever, and claim they have found a significant improvement for Chess Tiger.
It is a feature designed to please the ego of the user, who would think that he
is able to improve a chess engine without actually knowing anything about the
inner workings of that engine.
So you see, I do not have a great opinion of this feature, but I might add it if
it can help to sell more.
Sorry for being a little rude, I'm just trying to express my opinion honestly.
>Another silly question, do you think Tiger could be made to run on the latest
>HP-calculator, with ARM-processor? http://www.hpcalc.org/hp49gplus.php
>Just curious...
In theory I think it would be possible. In practice I think it would be hard,
depending on the tools HP would provide for developping software of it.
>Gambit Tiger 2.0 Aggressive does find Nxh6 quickly on my comp
>
>
>00:00:00.4 0,89 6 40180 Qd4 Nd3 Bxd3 cxd3 Rxd3 h5 Rb3 hxg4 Rxb7
>00:00:00.5 0,36 6 41252 Qd4 Nf4 Rge3 d5 Nef6 Rxe3 fxe3
>00:00:01.2 0,20 7 118568 Qd4 c3 Rxc3 Nxc3 Bxc3 d5 Bxb4 dxe4 Bc3
>00:00:02.1 -0,06 8 236752 Qd4 c3 Rxc3 Nxc3 Bxc3 d5 Bxb4 dxe4 Bc3 Rd6
>00:00:04.4 -0,06 9 543798 Qd4 c3 Rxc3
>00:00:09.9 -0,86 10 1282499 Qd4 c3 Bxc3 Nxc3 Qxb4 Nxe4 Qxb7 Rb8 Qxb8 Qxb8 Bxe4
>00:00:34.3 -0,86 11 4745038 Qd4 c3
>00:00:44.3 -1,20 12 6207282 Qd4 c3 Bxc3 Nxc3 Qxb4 Nxe4 Qxb7 Rb8 Qxb8 Qxb8 Bxe4
>d5 Rb1
>00:01:43.7 -0,66 12 15375286 Nxh6 c3 Nf5 cxb2 Qg4 Qf7 Re2 Be7 Nxe7 Qxe7 Rxb2 Re8
>Ng5
>00:04:37.6 -0,74 13 41862074 Nxh6 c3 Nf5 cxb2 Qg4 Rc8 Nxg7 Rg6 Nxe8 Rxg4 Rxg4
>Rxe8 Rg6
>00:06:32.7 -0,58 13 58828694 Ne3 Rxe4 Bxe4 Qxe4 Nxd5 Qxd5 Qc1 Qf5 Qxh6+ Qh7 Qe6
>Nd3 Re2 Nxb2 Rxb2
>00:08:46.0 -0,70 14 78188248 Ne3 Rxe4 Bxe4 Qxe4 Nxd5 Qxd5 Qc1 Qf5 Qxh6+ Qh7 Qg5
>Nd3 Re2 Nxb2 Rxb2
>00:14:19.3 -0,48 14 130894102 Nxh6 c3 Nf5 cxb2 Qg4 Rc8 Qh4+ Kg8 Kh2 Rc4 Nh6+
>Rxh6 Qxh6 Qf7 Rg4
>00:44:30.0 0,42 15 420013828 Nxh6
>
>
>Gambit Tiger 2.0 Aggressive Celeron 500Mhz 24 Mb hash
Chess Tiger X gambit aggressive style does better. P3M-933MHz, 6Mb HT (it is
running on Dell X200 laptop):
N8 0.43s Qd4 Nd3 Bxd3 cxd3 Rxd3 Nf4 Rde3 Qf7 Bc3 +0.52
N9 1.04s Qd4 c3 Rxc3 Nxc3 Bxc3 Na2 Bd2 h5 Nh2 +0.08
N10 2.36s Qd4 c3 Rxc3 Nxc3 Bxc3 Na2 Bd2 h5 Nh2 +0.08
N11 5.21s Qd4 c3 Rxc3 Nxc3 Bxc3 Na2 Bd2 h5 Nh2 d5 Qb2 -0.28
N12 15.76s Qd4 c3 Rxc3 Nxc3 Bxc3 d5 Bxb4 Qxa4 Nxh6 -0.16
N13 111.66s Nxh6 c3 Nf5 cxb2 Qg4 Qf7 Kh2 Kg8 Ng5 Rg6 Qh5 -0.06
N14 246.45s Nxh6 c3 Nf5 cxb2 Qg4 Qf7 Kh2 Kg8 Ng5 Rg6 Qh5 Nf6 -0.06
N15 945.10s Nxh6 c3 Nf5 cxb2 Qg4 Qf7 Re2 Be7 Nxe7 Nf4 Ng5 +0.84
N15 1381.15s Nxh6 c3 Nf5 cxb2 Qg4 Qd7 Kh2 g6 Nd4 Bc8 Nc5 +0.98
Nxh6 found in 1 minute 52 seconds and keeping it. A positive score (+0.84) found
in less than 16 minutes.
Chess Tiger X normal style is not impressive on this one...
N11 11.53s Qd4 c3 Bxc3 Nxc3 Qxb4 Nxe4 Qxb7 Rb8 Qxb8 Qxb8 -0.88
N12 23.83s Qd4 c3 Bxc3 Nxc3 Qxb4 Nxe4 Qxb7 Rb8 Qxb8 Qxb8 Bxe4 d5 -1.22
N12 35.37s Kf1 Nf4 Ngf6 Qf7 Rf3 Nbd3 Bxd3 Nxd3 Nh5 Qxh5 Rxf8+ Rxf8... -1.04
N13 111.22s Kf1 Nd3 Bxd3 cxd3 Nc3 Rxe1+ Qxe1 Qxe1+ Kxe1 Nf4 Ne3 Re8 -1.16
N13 170.15s Ne3 Rxe4 Bxe4 Qxe4 Nxd5 Qxd5 Qc1 Qf5 Qxh6+ Qh7 Qe6 -0.92
N14 242.38s Ne3 Rxe4 Bxe4 Qxe4 Nxd5 Qxd5 Qc1 Qh5 Qxc4 Nd5 Re6 Qd1+ -0.82
N15 519.75s Ne3 Rxe4 Bxe4 Qxe4 Nxd5 Qxd5 Qc1 Qf7 Qxh6+ -1.16
N15 2103.70s Nxh6 c3 Nf5 cxb2 Qg4 Qf7 Re2 Qc7 Ng5 Rh6 Nxh6 -0.70
N16 5089.94s Nxh6 c3 Nf5 cxb2 Qg4 Qd7 Qh5+ Kg8 Nf6+ Nxf6 -0.30
N1714129.96s Nxh6 c3 Nf5 cxb2 Qg4 Qd7 Kh2 g6 Nh4 Bc6 -0.28
So it needs 35 minutes to find Nxh6. I let it run overnight and it is still
thinking on Nxh6 at ply 18 but no score yet. So after 12 hours it still does not
have a positive score for Nxh6. I have just stopped it, the node counter was
close to 21 billion positions! It's one quarter of the number of nodes the
Diablo was able to compute in 14 months.
>I hope Chess Tiger X will Roar!
I'm doing everything to make it roar!
Christophe
This page took 0.01 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.