Your Country Needs You!

This forum is for general discussions and questions, including Collectors Corner and anything to do with Computer chess.

Moderators: Harvey Williamson, Steve B, Watchman

Forum rules
This textbox is used to restore diagrams posted with the fen tag before the upgrade.
Post Reply
User avatar
spacious_mind
Senior Member
Posts: 3999
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

I have added Sargon 5 playing with 386-25 MHz and 386-16 MHz. I also added Gavon Roconante v. 2.0. The updated chart is below.

Image

Best regards

Nick
Nick
User avatar
spacious_mind
Senior Member
Posts: 3999
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

Here are also some games I played. The first two games were Tasc CM 512K – 32MHz Gideon 3.1 Madrid with Normal style setting against HP VECTRA DX2-66MHz SARGON V. Both games were quite easy wins for the Chessmachine that showed in the first game its skill in pawn endings.

Game 1

HP VECTRA DX2–66 MHZ SARGON 5,LV 30S (2194) - TASC CM 512K-32 GIDEON 3.1,LV 30S (2430)
COMPUTER TEST MATCH Alabama, 15.03.2015

1.e4 Nf6 2.e5 Nd5 3.d4 d6 4.Nf3 g6 [HP VECTRA DX2–66 MHZ SARGON 5 OUT OF BOOK]

5.c4 [TASC CM 512K-32 GIDEON 3.1 OUT OF BOOK]

5...Nb6 6.Nc3 Bg7 7.Bd3 dxe5 8.dxe5 Bg4 9.Bf4 Nc6 10.Be2 Bxf3 11.Bxf3 Nxc4 12.Bxc6+ bxc6 13.Qa4 Nxe5 14.Rd1 Qc8 15.Bxe5 Bxe5 16.Qxc6+ Kf8 17.0–0 Rb8 18.f4 Bd6 19.b3 Qb7 20.Qc4 Qb4 21.Qxb4 Rxb4 22.g3 Kg7 23.Kf2 Rbb8 24.Ke3 Rhd8 25.Rd3 Bc5+ 26.Ke2 Rd6 27.Rfd1 Rbd8 28.Nb5 Rxd3 29.Rxd3 Rxd3 30.Kxd3 Bb6 31.Nd4 Bxd4 32.Kxd4 Kf6 33.Ke4 g5 34.fxg5+ Kxg5 35.h3 h5 36.b4 f5+ 37.Ke5 f4 38.gxf4+ Kh4 39.Ke4 Kxh3 40.Kf3 Kh4 41.b5 e6 42.a3 Kh3 43.a4 Kh4 44.a5 Kh3 45.a6

0–1

Game 2

Game 2 was quite an easy win for Gideon 3.1 Madrid against Sargon 5.

TASC CM 512K-32 GIDEON 3.1,LV 30S (2430) - HP VECTRA DX2–66 MHZ SARGON 5,LV 30S (2194)
COMPUTER TEST MATCH Alabama, 15.03.2015

1.c4 e5 2.Nc3 Nf6 3.Nf3 [HP VECTRA DX2–66 MHZ SARGON 5 OUT OF BOOK]

3...Nc6 4.g3 d5 5.cxd5 Nxd5 6.Bg2 Nxc3 [TASC CM 512K-32 GIDEON 3.1 OUT OF BOOK]

7.bxc3 e4 8.Ng1 Bf5 9.Rb1 Rb8 10.Qa4 Qf6 11.Bxe4 b5 12.Qc2 Ne7 13.Bxf5 Nxf5 14.Nf3 Be7 15.c4 c6 16.Bb2 Qg6 17.cxb5 cxb5 18.Be5 Bd6 19.Qc6+ Kf8 20.Rxb5 Rxb5 21.Qxb5 Ne7 22.0–0 Qe6 23.Rb1 Bxe5 24.Qxe5 Qxe5 25.Nxe5 g6 26.Rb8+ Kg7 27.Rxh8 Kxh8 28.Nxf7+ Kg8 29.Ne5 Kf8 30.e4 Nc8 31.d4 Ke7 32.Nc4 Nb6 33.Nxb6 axb6 34.f4

1–0

Game 3

Next I tried Gideon 3.1 Madrid against the highest test rated program so far which is Gavon CPW v.1.1. This was a good game both were quite evenly matched.

[Event "COMPUTER TEST MATCH"]
[Site "Alabama"]
[Date "2015.03.15"]
[Round "?"]
[White "GAVON CPW V.1.1, LV AT30."]
[Black "TASC CM 512K-32 GIDEON 3.1, LV 30S."]
[Result "1/2-1/2"]
[WhiteElo "2498"]
[BlackElo "2430"]
[PlyCount "103"]
[EventDate "2015.03.15"]
[EventCountry "USA"]

1. e4 Nf6 2. e5 Nd5 3. d4 d6 4. Nf3 Bg4 5. Be2 c6 {CPW V.1.1 OUT OF BOOK} 6.
O-O {TASC CM 512K-32 GIDEON 3.1 OUT OF BOOK} dxe5 7. Nxe5 Bxe2 8. Qxe2 e6 9.
Qh5 g6 10. Qe2 Bg7 11. Rd1 O-O 12. c4 Ne7 13. Nc3 Nf5 14. Bf4 Nd7 15. Ne4 Rc8
16. g4 Nxe5 17. dxe5 Nd4 18. Qe3 c5 19. Nxc5 Rxc5 20. Rxd4 Qc7 21. Rd7 Qxd7 22.
Qxc5 Rc8 23. Qxa7 Rxc4 24. Qe3 Qd4 25. Qxd4 Rxd4 26. Bg3 Rxg4 27. f4 g5 28. Rf1
f5 29. exf6 gxf4 30. Rxf4 Rxf4 31. Bxf4 Bxf6 32. Bc1 Kf7 33. Kf2 Kg6 34. b3 Kf5
35. Ke2 Kg4 36. Kf2 Kh3 37. Bf4 h5 38. Kf3 e5 39. Bd2 Kxh2 40. Ke4 Kg2 41. Be1
h4 42. Bxh4 Bxh4 43. Kxe5 Kf3 44. Kd6 Bg3+ 45. Kd7 Ke4 46. a4 Kd5 47. a5 Bd6
48. b4 Bxb4 49. Kc7 b5 50. axb6 Ba5 51. Kb7 Bxb6 52. Kxb6 1/2-1/2

Game 4

Lastly I decided to play Saitek Analyst D with 6 MHz against the 486-66 Mhz Sargon V. Analyst D should have won this game. It was amazing that it stumbled right at the end with a clear win just about a move away.

HP VECTRA DX2–66 MHZ SARGON 5,LV 30S (2194) - SAITEK ANALYST D - 6 MHZ,LV E3 (2116)
COMPUTER TEST MATCH Alabama, 15.03.2015

1.e4 e6 2.d4 [HP VECTRA DX2–66 MHZ SARGON 5 OUT OF BOOK]

2...d5 3.Nd2 c5 4.exd5 Qxd5 5.Ngf3 cxd4 6.Bc4 Qc5 [SAITEK ANALYST D - 6 MHZ OUT OF BOOK]

7.Qe2 Nc6 8.Ne4 Qb4+ 9.c3 Qb6 10.Nxd4 Nxd4 11.cxd4 Qxd4 12.0–0 Bd7 13.Rd1 Qe5 14.f4 Qa5 15.Be3 Rd8 16.Rac1 Nh6 17.Bd4 Nf5 18.Bf2 Be7 19.g4 Nh6 20.Rxd7 Rxd7 21.Bb5 a6 22.Rc8+ Bd8 23.Nd6+ Ke7 24.Bxd7 Kxd7 25.Nxb7 Qxa2 26.Ra8 Qb1+ 27.Kg2 Qa2 28.g5 Qd5+ 29.Kg1 Nf5 30.Qxa6 Qd1+ 31.Qf1 Qg4+ 32.Kh1 Qf3+ 33.Qg2 Qxf4 34.Nxd8 Qc1+ 35.Bg1 Nd6 36.Nxf7 Rxa8 37.Ne5+ Ke7 38.Qxa8 Qxg5 39.Qa7+ Kf6 40.Nf3 Qd5 41.Bd4+ Kf5 42.Kg2 Nb7 43.Qb8 g5 44.Qh8 Kf4 45.Qf6+ Ke4 46.Bf2 g4 47.Ng5+ Kd3+ 48.Kg3 e5 49.Kxg4 Kc2 50.Bg3 h6 51.Nf7 e4 52.Qc3+ Kb1 53.Nxh6 Qd1+ 54.Kg5 Qd8+ 55.Kh5 Qa5+ 56.Qxa5 Nxa5 57.b4 Nc4 58.h4 Kc2 59.Kg4 e3 60.Kf3 Kd3 61.Ng4 Nd2+ 62.Kg2 e2 63.h5 Kc2 64.h6 Ne4 65.Ne3+ Kd3 66.h7 Nxg3 67.Kf2 Ne4+ 68.Ke1 Ng5 69.Ng2 Nxh7 70.b5 Nf6 71.b6 Nd7 72.b7 Kd4 73.Nf4 Kc5 74.Ng6

½–½


[fen]8/8/7P/8/1P2n3/3kN1B1/4p1K1/8 w - - 0 66[/fen]

In this position Saitek Analyst D 6 MHz played 66. h7?? which gave away the win. 66. Be1 would have easily won. It was just too easy to miss. Amazing blunder.

Off to see Santana tonight in Concert regards,
Nick
User avatar
Steve B
Site Admin
Posts: 10140
Joined: Sun Jul 29, 2007 10:02 am
Location: New York City USofA
Contact:

Post by Steve B »

spacious_mind wrote:

Off to see Santana tonight in Concert regards,

Enjoy it!

Oye Como Va Regards
Steve
User avatar
paulwise3
Senior Member
Posts: 1505
Joined: Tue Jan 06, 2015 10:56 am
Location: Eindhoven, Netherlands

Post by paulwise3 »

spacious_mind wrote:In this position Saitek Analyst D 6 MHz played 66. h7?? which gave away the win. 66. Be1 would have easily won. It was just too easy to miss. Amazing blunder.
Hi Nick, computing just one ply deeper or not, can make a great difference. I found out before. To see this is wrong, it has to look at least 7 or 8 plies deep, which depending on the given time may have been a (ply-)bridge too far.

Deep ply regards,
Paul.
User avatar
paulwise3
Senior Member
Posts: 1505
Joined: Tue Jan 06, 2015 10:56 am
Location: Eindhoven, Netherlands

Post by paulwise3 »

Just got a Team Mate in the house. My very first machine was a Tandy 1850 (picked it up for 2,50 dutch guilders ~~ 1,15 euro). I noticed that the specs and programmers were exactly the same, so did a quick test with testgame 3: only one move was different! (4.Nc3 instead of 4.Qg4). So it should be safe to conclude they are clones. Below are the details:

Code: Select all

COMPUTER TEST RESULT	COMPUTER TEST RESULT
Saitek	  Tandy
Team Mate	  1850
Julio Kaplan, Barnes, Craig	Julio Kaplan, Barnes, Craig
Level 6 (30 seconds) average time	Level 4 (30 seconds) average time
6301Y  - 8 Bit - 2/8 MHz - 16 KB ROM - 256 Byte RAM	6301Y  - 8 Bit - 2/8 MHz - 16 KB ROM - 256 Byte RAM
WHITE	WHITE
4.Qg4	4.Nc3
5.Nf3	        5.Nf3
6.Bxf4	6.Bxf4
7.e5	        7.e5
8.Bxf7+	8.Bxf7+
9.Bxg8	9.Bxg8
10.Nec3	10.Nec3
11.Kc1	11.Kc1
12.Nxc3	12.Nxc3
13.Qg4+	13.Qg4+
14.Qxg7	14.Qxg7
15.Bxg8	15.Bxg8
16.Bxd6+	16.Bxd6+
17.Bxd6+	17.Bxd6+
18.Bd5+	18.Bd5+
19.d4	19.d4
20.Bxa8	20.Bxa8
21.Bd6	21.Bd6
22.Qc3+	22.Qc3+
WHITE SCORE:	WHITE SCORE:
2033	       1929
BLACK MOVES	BLACK MOVES
3. ... Nf6	 3. ... Nf6
4. ... d6	 4. ... d6
5. ... Qb6	 5. ... Qb6
6. ... Qb6	 6. ... Qb6
7. ... Qxb2	 7. ... Qxb2
8. ... Kf8	 8. ... Kf8
9. ... Qxa1	 9. ... Qxa1
10. ... Nf6	 10. ... Nf6
11. ... Bxc3+	11. ... Bxc3+
12. ... Qxh1	12. ... Qxh1
13. ... Kc7	        13. ... Kc7
14. ... Nh6 	14. ... Nh6
15. ... Ngf6	15. ... Ngf6
16. ... cxb5	16. ... cxb5
17. ... Kd8 	17. ... Kd8
18. ... Ka6 	18. ... Ka6
19. ... b4   	19. ... b4
20. ... Qxg2+	20. ... Qxg2+
21. ... Kxb4	21. ... Kxb4
BLACK SCORE:	BLACK SCORE:
1721         	1721
TOTAL SCORE:	TOTAL SCORE:
1877         	1825
User avatar
spacious_mind
Senior Member
Posts: 3999
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

paulwise3 wrote:Just got a Team Mate in the house. My very first machine was a Tandy 1850 (picked it up for 2,50 dutch guilders ~~ 1,15 euro). I noticed that the specs and programmers were exactly the same, so did a quick test with testgame 3: only one move was different! (4.Nc3 instead of 4.Qg4). So it should be safe to conclude they are clones. Below are the details:
Hi Paul,

Playing game 3 is a good choice for a quick clone test as in game 1 too many computers follow the same moves.

I am next trying out RexChess from Don Dailey and Larry Kaufman. Rexchess has a built in rating to calculate it's playing strength. Which is as follows:

Magnavox Maxstation 386-16 MHz Desktop = Rexchess rating of 2198
Dell 325NC 386-25 MHz Laptop = 2271
Compaq Contura 486-25 MHz Laptop = 2381
HP Vectra 486-66 MHz Desktop = 2504

So I will be curious to see how the test rating compares to this.

Best regards
Nick
User avatar
spacious_mind
Senior Member
Posts: 3999
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

Steve B wrote:
spacious_mind wrote:

Off to see Santana tonight in Concert regards,

Enjoy it!

Oye Como Va Regards
Steve
Thanks Steve, it was a great concert, very loud my ears a still ringing. They did play Oye Como Va :P

Best regards
Nick
User avatar
paulwise3
Senior Member
Posts: 1505
Joined: Tue Jan 06, 2015 10:56 am
Location: Eindhoven, Netherlands

Post by paulwise3 »

Hi Nick,
I finally got my Excalibur Grandmaster. What a comfortable pleasure to play with! I started with the testgames, and noticed some slight deviations from your Platinum, which surprises me a little. Will go into detail after the weekend, when I finished all five ;-)

GM regards,
Paul.
User avatar
spacious_mind
Senior Member
Posts: 3999
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

paulwise3 wrote:Hi Nick,
I finally got my Excalibur Grandmaster. What a comfortable pleasure to play with! I started with the testgames, and noticed some slight deviations from your Platinum, which surprises me a little. Will go into detail after the weekend, when I finished all five ;-)

GM regards,
Paul.
Hi Paul,

I have both models so I can compare mine sometime to see if there really is a difference. The Excalibur GM score is what Steve had tested with his a while back.

Regards
Nick
User avatar
spacious_mind
Senior Member
Posts: 3999
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

I have added another Gavon to the Test, Bismarck 1.2 written by Evgeny Shtranvasser. It scored quite well finishing with a final rating of 2164 ELO.

So far I have tested the following six Gavon chess programs:

GAVON CHESS PROGRAMS WITH COMPLETED TEST SCORE

1 CPW V.1.1 2498 ELO
2 BISMARCK 1.2 2164 ELO
3 REDQUEEN V. 0.4 1972 ELO
4 ROCINANTE V. 2.0 1952 ELO
5 ROCINANTE V. 1.01 1865 ELO
6 JFRESH V.0.1A 1854 ELO

Mapping it into a chart, it looks as follows:

GAVON CHESS PROGRAMS FINAL RATING AFTER 5 TEST GAMES

Image

With Gavon there are another 8 programs lower rated than CPW v. 1.1 so it will be interesting to me to progressively track these as well on a chart.

I am also using these tests to closely follow the rating progression of two DOS programs starting from a lowest possible Hardware which is a 286-10 MHz Desktop computer that I have all the way through to as fast as I can get it.

The two programs that I will test are:

SARGON 5
REXCHESS 2.30

Rexchess is interesting because it has a built in rating system that estimates its strength on different hardware and also provides the nodes, which in the end will be great to compare. Other DOS software that provides ratings are the Rebel programs from Ed Schroeder. Problem with these is that they don't play down to a 286, therefore they are slightly less interesting for my test.

With Sargon 5 I have completed the tests on the following hardware already:

CURRENT DOS HARDWARE WITH COMPLETED TEST SCORE FOR SARGON 5

1 HP VECTRA 486DX2-66MHZ 2194 ELO
2 AST ADVANTAGE! NB 486SX-33 MHZ 2134 ELO
3 COMPAQ CONTURA 486-25 MHZ 2103 ELO
4 AST PREMIUM EXEC 386SX-25 MHZ 2027 ELO
5 DELL 325NC 386SL-16 MHZ 2094 ELO
6 MAXSTATION 386SX-16 MHZ 1979 ELO
7 COMPAQ SLT/286-12 MHZ 1948 ELO

I also mapped Sargon 5 on a Chart so that I could follow the progress:

DOS HARDWARE WITH COMPLETED TESTS FOR SARGON 5

Image

The only surprise is Dell's computer Model 325NC-386SL-16MHz. That must have been a good computer in it's day for chess. All the other hardware so far follows the expected hardware improvement progression.

I have several more old computers to test with Sargon 5 and Rexchess 2.30 which I will report on another day.

Below is the rating chart for all the currently tested programs. 28 in total so far!

RATING CHART FOR PROGRAMS THAT HAVE COMPLETED THE 5 TEST GAMES.

Image

@Fernando, you mentioned that you are playing Sargon 5 on a 3 GHz computer. Are you using DOSbox? I am curious because under DOSbox, there would be a considerable speed loss.

Best regards
Nick
User avatar
Bryan Whitby
Senior Member
Posts: 1001
Joined: Wed Feb 18, 2009 9:57 pm
Location: England

Post by Bryan Whitby »

Hi Nick

Great that you are including Gavon engines in the test and Crafty can also now be tested with Josu releasing it at the weekend.

Keep up the great work
Bryan
mychess
Member
Posts: 96
Joined: Tue Jul 31, 2007 8:55 pm

Post by mychess »

Hello,

The i386sl contained support for an external cache of 16 to 64 kB.
For the Dell 325n/NC, the cache is 64 KB, 4-way set associative.
So it had, for chess programs, the same power than a i486 at the same frequency.

Sargon V is , IMHO, the program of the Excel 68000 Mach II.


Olivier
User avatar
spacious_mind
Senior Member
Posts: 3999
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

mychess wrote:Hello,

The i386sl contained support for an external cache of 16 to 64 kB.
For the Dell 325n/NC, the cache is 64 KB, 4-way set associative.
So it had, for chess programs, the same power than a i486 at the same frequency.

Sargon V is , IMHO, the program of the Excel 68000 Mach II.


Olivier
Hi Olivier,

Thanks, that would explain why the Dell computer performed better than the other 386 computers.

Nick
Nick
User avatar
spacious_mind
Senior Member
Posts: 3999
Joined: Wed Aug 01, 2007 10:20 pm
Location: Alabama
Contact:

Post by spacious_mind »

I added 3 more computers to the Sargon 5 test. All 3 are Pentiums, 75 MHz, 100 Mhz and 120 Mhz.

1 COMPAQ ARMADA 4120 PENTIUM 120 MHZ 2280
2 TOSHIBA SATELLITE 100CS PENTIUM 75 MHZ 2261
3 FUJISTSU MILAN PENTIUM 100 MHZ 2238
4 HP VECTRA 486DX2-66MHZ 2194
5 AST ADVANTAGE! NB 486SX-33 MHZ 2134
6 COMPAQ CONTURA 486-25 MHZ 2103
7 DELL 325NC 386SL-16 MHZ 2094
8 AST PREMIUM EXEC 386SX-25 MHZ 2027
9 MAXSTATION 386SX-16 MHZ 1979
10 COMPAQ SLT/286-12 MHZ 1948

Image

The Toshiba Satellite 75 MHz outperforms the Fujitsu Milan 100 MHz. Just like the Dell 325NC this one must also have been superior in it's day.

The ELO has grown by 332 ELO starting with the Compaq SLT/286-12 MHz and finishing with Compaq Armada 4120 Pentium 120 MHz.

Best Regards
Nick
User avatar
paulwise3
Senior Member
Posts: 1505
Joined: Tue Jan 06, 2015 10:56 am
Location: Eindhoven, Netherlands

Post by paulwise3 »

Hi Nick,
I finished the 5 testgames for the Excalibur Grandmaster (running on alkaline batteries). It scored better then Steve's Platinum. I wonder if Steve tested it with 30s fixed instead of average, or could there really be a difference in playing style/strength? The playing style looked the same though.

The first test was with game 3, and as I just started using the GM, I had some fights with it taking back moves and executing them again. Sometimes it got confused, and I had to start all over. So finally I decided not to take back the moves that it played the same as in the testgame. And I continued doing that in the other testgames. With testgame 5, at move 20 I forgot to play the right move for black, detected it only 3 moves later. Not willing to start all over again, I took back those moves, then played the correct ones. Just hope it didn't influence its time management too much. Anyway, that was the game it scored worst.

Here are the global results, I will mail you the detailed moves.
(W-White elo, B-Black elo, T-Total elo)

Code: Select all

 test01 test02 test03 test04 test05 (1..5)/5
W-2220	2341	2251	2083	1937	2166
B-2164	1881	1836	2026	1741	1930
T-2192	2111	2044	2054	1839	2048
So it's average score comes close to the Sphinx Legend/Concerto.

Elo regards,
Paul.
Post Reply