WEAK ENGINES????

[spacious_mind]

Trying Tibono's suggestion of Dynamic and Max, Complete Chess System found the solution in 14 seconds. Therefore approx. 40% slower than Amiga 68060.

Well it seems Dynamic and Max is about 3-4 times faster than Auto and Max. Chessmaster 2100 just reached 3.3 Million positions compared to 750K positions with Auto and Max.

Regards

[Martin Hertz]

Complete Chess System without hash (woks with EMS only) needs 4 seconds.
CC2175 needs 12 seconds without hash and 3 seconds with 8 MB hash.
The 680000 versions of some programs are very effecient.

[Fernando]

Martin,

I just tried Complete Chess Systems. I take it back they do seem to be quite similar.

The Test Position D-Fend at 100% auto takes about 30 seconds to find the mate.

Amiga 68060 takes about 10 seconds to find the mate.

Based on this your computer at 3.3 GHz should be very quick. Please let me know how it does Complete Chess Systems.

Best regards

In D-Fedn Turbochess goes from around 800 n/s with 10% to near 30 thousand with 99% in 30 seconds.

technical regards
F

[spacious_mind]

Martin,

I just tried Complete Chess Systems. I take it back they do seem to be quite similar.

The Test Position D-Fend at 100% auto takes about 30 seconds to find the mate.

Amiga 68060 takes about 10 seconds to find the mate.

Based on this your computer at 3.3 GHz should be very quick. Please let me know how it does Complete Chess Systems.

Best regards

In D-Fedn Turbochess goes from around 800 n/s with 10% to near 30 thousand with 99% in 30 seconds.

technical regards
F

Yes with Auto and Max it did about 10,000 n/s and with Dynamic and Max it does 30,245 n/s.

I tested it against Amiga CP2150 and it had a tough time with setting Auto and Max. It got smashed.

[Event "Computer Test Match"]
[Site "?"]
[Date "2017.02.19"]
[Round "2"]
[White "DB Turbo Chess, 30S AV.."]
[Black "AM68060 CP 2150, 30S AV.."]
[Result "0-1"]
[ECO "D50"]
[WhiteElo "2107"]
[BlackElo "2100"]
[Annotator "SM"]
[PlyCount "50"]
[EventDate "2017.02.19"]
[EventType "match"]
[EventRounds "2"]

1. d4 d5 2. c4 e6 3. Nc3 Nf6 4. Bg5 c5 5. cxd5 Qb6 {DB Turbo Chess out of book} 6. Bxf6 gxf6 7. e4 {AM68060 CP 2150 out of book} cxd4 8. Qa4+ Bd7 9. Nb5 Na6 10. Qxd4 Bb4+ 11. Kd1 Bc5 12. Qxf6 O-O 13. dxe6 fxe6 14. Qg5+ Kh8 15. Qe5+ Kg8 16. Qg3+ Kf7 17. Qd3 Rad8 18. Kc1 Bc6 19. Qe2 Bxf2 20. Qh5+ Ke7 21. b3 Be3+ 22. Kb1 Rxf1+ 23. Kb2 Rd2+ 24. Kc3 Qa5+ 25. Kc4 Qb4# 0-1

Will trying it again with Dynamic + Max where it reaches 8 ply at 30 seconds setting. With Auto + Max only 6 ply.



It is a very colorful program and seems to play like it too. The above nodes are from the Dynamic + Max setting.

Best regards

[Martin Hertz]

I've played many games at 3.3 GHz with Turbo-Chess and CP2150. CP2150 is clearly stronger, but sometimes
CP2150 loses to Turbo-Chess because of a Shannon B blunder, whats typical for the CP2150, CC2175 and CCS.
I haven't observed this type of blunders in Chess System Tal DOS. So I think CST is not really a Shannon B
program anymore but a program with very special evaluation. CST is much stronger than it's predecessors.

The DOS emulation at it's best is still around 25 times slower than real DOS. This means normally 2 plies
less in the same time. This makes a significant difference in the playing strength, especially in tactics.

[spacious_mind]

I've played many games at 3.3 GHz with Turbo-Chess and CP2150. CP2150 is clearly stronger, but sometimes
CP2150 loses to Turbo-Chess because of a Shannon B blunder, whats typical for the CP2150, CC2175 and CCS.
I haven't observed this type of blunders in Chess System Tal DOS. So I think CST is not really a Shannon B
program anymore but a program with very special evaluation. CST is much stronger than it's predecessors.

The DOS emulation at it's best is still around 25 times slower than real DOS. This means normally 2 plies
less in the same time. This makes a significant difference in the playing strength, especially in tactics.

So what did you do, get yourself a 3.3 GHz computer and loaded DOS on it?

Regards

[spacious_mind]

I did play a second game, this time against Amiga Chess Champion 2175 with Dynamic and Max setup.

[Event "Computer Test Match"]
[Site "?"]
[Date "2017"]
[Round "2"]
[White "AM CC2175, 30S AV.."]
[Black "DB Turbo Chess, 30S AV.."]
[Result "1/2-1/2"]
[ECO "D50"]
[WhiteElo "2175"]
[BlackElo "2100"]
[Annotator "SM"]
[PlyCount "110"]
[EventDate "2017"]
[EventType "match"]
[EventRounds "2"]

1. Nf3 Nf6 2. d4 {DB Turbo Chess out of book} Nc6 {AM CC2175 out of book} 3. d5 Nb4 4. c4 c6 5. a3 Na6 6. Nc3 Nc5 7. b4 Nce4 8. Nxe4 Nxe4 9. dxc6 dxc6 10. Qxd8+ Kxd8 11. Ng5 Nxg5 12. Bxg5 Bf5 13. Rd1+ Ke8 14. e3 e5 15. Be2 a5 16. c5 f6 17. Bh4 axb4 18. axb4 Ra4 19. Bh5+ Bg6 20. Bxg6+ hxg6 21. Bg3 Rxb4 22. O-O Bxc5 23. Rd2 Ke7 24. Rfd1 Rg8 25. Rd7+ Ke6 26. Kf1 e4 27. Bd6 Bxd6 28. R1xd6+ Ke5 29. Rd1 Kf5 30. h3 g5 31. g4+ Ke6 32. R1d6+ Ke5 33. Rd1 f5 34. Re7+ Kf6 35. Rdd7 fxg4 36. hxg4 b5 37. Rf7+ Kg6 38. Rf5 Rc4 39. Rd6+ Kh7 40. Rxg5 g6 41. Rd4
Rxd4 42. exd4 Rd8 43. Rc5 Rxd4 44. Rxc6 Rc4 45. Rb6 b4 46. g5 Kg7 47. Ke2 Kf8 48. Ke3 Kf7 49. Kf4 e3+ 50. Kxe3 Rc3+ 51. Kd4 Rf3 52. Rxb4 Rf4+ 53. Kc5 Rxf2 54. Rb7+ Ke6 55. Rb6+ Kf7 1/2-1/2

Final Postion

[fen]8/5k2/1R4p1/2K3P1/8/8/5r2/8 w - - 0 56[/fen]

Turbo-Chess went up by two pawns but from that point onwards CC2175 outplayed it.

But, for a program that was written in 1985, it is really surprisingly good and has an interesting style of play.

Pity that I only just received the program as I think Turbo-Chess would be competitive in my Division 4 league. I might include it later in Division 3 but I fear that Division 3 might become a bit too much of a challenge for it. But we will see.

Best regards

[Martin Hertz]

Turbo-Chess is one of my DOS favorites, because it is well programmed, what is not always the case, especially
with very old programs. The evaluation is simple, but it is a good fighter.

The DOS emulation at it's best is still around 25 times slower than real DOS. This means normally 2 plies
less in the same time. This makes a significant difference in the playing strength, especially in tactics.

So what did you do, get yourself a 3.3 GHz computer and loaded DOS on it?

Yes, I use a notebook dedicated to DOS chess only. If you want to use an old notebook for this purpose, take at
least a Pentium M (Dothan) because it is cheap and very speed and energy efficient for DOS chess. The later intels
are only faster propotional to their clockrate compared to the Pentium M. The Pentium 3 is slow and the Pentium 4
has a very high TDP, what means it's energy wasting.

[spacious_mind]

Turbo-Chess is one of my DOS favorites, because it is well programmed, what is not always the case, especially
with very old programs. The evaluation is simple, but it is a good fighter.

The DOS emulation at it's best is still around 25 times slower than real DOS. This means normally 2 plies
less in the same time. This makes a significant difference in the playing strength, especially in tactics.

So what did you do, get yourself a 3.3 GHz computer and loaded DOS on it?

Yes, I use a notebook dedicated to DOS chess only. If you want to use an old notebook for this purpose, take at
least a Pentium M (Dothan) because it is cheap and very speed and energy efficient for DOS chess. The later intels
are only faster propotional to their clockrate compared to the Pentium M. The Pentium 3 is slow and the Pentium 4
has a very high TDP, what means it's energy wasting.

Thanks, just wondered what you use. I have a set of oldie laptops and desktops which I use only for DOS but these are more of the vintage range 286, 386, 486. I use these for example if I want to play two 286 programs against each other and I also have them at different speeds from 16 MHz to 133 MHz. I also have a couple of Pentium 300 for the same reason.

Best regards

[mclane]

I've played many games at 3.3 GHz with Turbo-Chess and CP2150. CP2150 is clearly stronger, but sometimes
CP2150 loses to Turbo-Chess because of a Shannon B blunder, whats typical for the CP2150, CC2175 and CCS.
I haven't observed this type of blunders in Chess System Tal DOS. So I think CST is not really a Shannon B
program anymore but a program with very special evaluation. CST is much stronger than it's predecessors.

The DOS emulation at it's best is still around 25 times slower than real DOS. This means normally 2 plies
less in the same time. This makes a significant difference in the playing strength, especially in tactics.

CSTAL prunes for interest. It extends moves that are interesting and prunes those of no interest. This leads to trees bigger then nullmove trees.
But CSTAL can drive into these noisy areas of the tree because it has special knowledge. The other programs die in those noisy areas. Also humans.
Humans prefer silent positions. Where they have overview. They fail in chaotic positions.

[Martin Hertz]

Thanks Thorsten for the explanation. I like the Shannon B programs and it's fantastic to see them play at nowadays speed.
I've not that much experience with CC2175 but the CP2150 finds very often good moves or avoid mistakes in ply 6 or 7.
A game with ply 5 or less is much weaker. This effect seems much stronger compared to the Shannon A or mixed type programs.
CC2175 goes much deeper in the same time, but in most cases it needs more plies than the CP2150 to find a move. I'll make
further tests with CC2175 at max speed. It would be interesting to see, how the Mephisto III program plays at that speed.

[mclane]

I want to come back to this old thread.
I hope all participants are still alive and member of this forum.

I have clockchess89 for zx spectrum now in a dedicated chess computer tournament, slightly oc with factor 3.

In the same tournament i have colossus chess participating on 5 mhz instead the normal 1 mhz in the c64 machine.

By putting clockchess to 10 mhz i thought it has equal hardware then colossus chess.

Imo it looks to me like superchess3.5 is a predecessor of clockchess89 but has no permanent brain.
And permanent brain is a feature that is imo important in 40/120 games.

I would have put an early richard lang (cyrus) in the tournament, or even as dedicated chess computer, but as far as i see it, all software cyrus versions have no permanent brain. Also the hanimex hardware cyrus machines on z80 have no permanent brain,
I think even the very expensive l’empereur has no permanent brain (or can anybody present evidence that i am wrong ?!).

The tournament can be found here:
https://www.schachcomputer.info/forum/s ... 441&page=4