Rom Nelsons dedicated chess computers

mclane · Post by **mclane** » Sun Apr 26, 2015 12:11 pm

I do have a Sensory voice chess challenger here and my friend has the challenger 10.

We will play some games to show how Ron Nelsons programs play.
We can then extrapolate how this program would play on a far superior hardware like the h8 CPU with 12 or 24 MHz.

Mostly the opponents have 6502 CPUs. The 6502 is faster than the Z80 CPU. I once read the 6502 is 2 times faster if it has similar MHz.

So this is a little disadvantage when playing against
Chess champion super system III or intelligent chess or other opponents.

Challenger 10 has 1153 Elo, following schachcomputer.info.
sensory voice chess challenger has 1394 Elo following schachcomputer.info.

Chess champion super system III has 1375.
Intelligent chess has 1100.

Of course we cannot be sure that these Elo are right. We have to make our own conclusions.
Btw, spracklens sensory voice chess challenger champion has 1550 Elo.
That's 150 Elo difference from Ron Nelson to spracklens.
We see that z80 is 1/2 speed of 6502. So we can say the spracklens software
Was 100 Elo stronger. But there is another problem,
Spracklens program has permanent brain, Ron Nelsons early creatures do not have permanent brain. How much Elo is permanent brain ?
I would say 50.

ricard60 · Post by **ricard60** » Sun Apr 26, 2015 1:31 pm

About 6502 CPUs faster than Z-80 thats right, this have to deal with the internal cycles of each CPU. It will be very interesting to see as how Ron Nelson program plays on a superior hardware.

mclane · Post by **mclane** » Sun Apr 26, 2015 4:11 pm

[Event "40/120"]
[Site "SCW"]
[Date "2015.04.11"]
[Round "1"]
[White "CChampion Super System III"]
[Black "Fidelity Challenger 10 B"]
[ECO "C13"]
[Result "1/2-1/2"]

1. e4 e6 2. d4 d5 3. Nc3 Nf6 4. Bg5 Be7 5. exd5 Nxd5
6. Bxe7 Qxe7 7. Nxd5 exd5+ 8. Qe2 Nc6 9. Qxe7+ Nxe7
10. O-O-O Bf5 11. Nf3 O-O 12. Bd3 Rae8 13. Bxf5 Nxf5
14. Rhe1 a5 15. Rxe8 Rxe8 16. a4 f6 17. h4 Nd6 18. h5 b6
19. g4 Re4 20. Rg1 Rf4 21. Rg3 Ne4 22. Rh3 c5 23. dxc5 Nxf2
24. Rg3 bxc5 25. b3 Ne4 26. Rh3 Rxg4 27. h6 gxh6 28. Rxh6
Rg3 29. Ne1 c4 30. bxc4 Rg1 31. Kd1 dxc4 32. Rh5 Ng5
33. Kd2 c3+ 34. Ke2 Rg4 35. Kd3 Rg1 1/2-1/2

the game does not look like MK3 beeing superior.

[Event "40/120"]
[Site "SCW"]
[Date "2015.04.26"]
[Round "1"]
[White "Intelligent Chess"]
[Black "Fidelity Chess Challenger 10"]
[ECO "A10"]
[Result "1/2-1/2"]

1. c4 d5 {"} 2. e3 {"} Nf6 3. Nc3 Bg4 4. f3 Be6 5. cxd5
Nxd5 6. Nxd5 Qxd5 7. d4 Nc6 8. Bd3 g6 9. Ne2 Bg7 10. O-O h5
11. Nf4 Qd6 12. Nxe6 Qxe6 13. Qb3 Qxb3 14. axb3 O-O-O
15. f4 h4 16. Bd2 e6 17. h3 Rh5 18. Bc3 Bf6 19. Rf2 Rhd5
20. Be4 Rb5 21. b4 Nxb4 22. Rxa7 Kb8 23. Rxb7+ Rxb7
24. Bxb7 Nd3 25. Rd2 Nxf4 26. Bc6 Rd6 27. Be8 Nd5 28. Bxf7
Bg5 29. Re2 Bxe3+ 30. Kf1 g5 31. b3 Rb6 32. b4 Bc1 33. Be1
Be3 34. Bg8 Rc6 35. b5 Rb6 36. Bf2 Bxf2 37. Kxf2 Nf4
38. Re3 Rxb5 39. Bxe6 Rb2+ 40. Kf3 Nxg2 41. Rb3+ Rxb3+
42. Bxb3 Nf4 43. Kg4 Ne2 44. d5 Nd4 45. Bc4 Kb7 46. Kxg5
1/2-1/2

Post by **Steve B** » Sun Apr 26, 2015 9:54 pm

Good Stuff Thor
Always enjoy seeing the really early computers play

Novag MKII Regards
Steve

afos99 · Post by **afos99** » Sun Jul 19, 2015 10:04 am

I have a few early chess challengers and am interested in various aspects around the history, performance etc. Recently found a page by Aart Bik telling how as a kid he found a way to beat his Sensory 8 on level one. I tried it on my version and it was the same. I then tried the same on a CC7 with the same result. Next I tried it on my VCC. As the VCC was supposed to be very similar to the CC7 except the addition of the book openings I thought it would produce the same result. However it didn't fall into the same trap because it used more computing time and obviously saw ahead.
My Sensory 9 did not get caught and gave more varied responses to not go the same route.

Dave

Post by **Fernando** » Sun Jul 19, 2015 3:07 pm

afos99 wrote:I have a few early chess challengers and am interested in various aspects around the history, performance etc. Recently found a page by Aart Bik telling how as a kid he found a way to beat his Sensory 8 on level one. I tried it on my version and it was the same. I then tried the same on a CC7 with the same result. Next I tried it on my VCC. As the VCC was supposed to be very similar to the CC7 except the addition of the book openings I thought it would produce the same result. However it didn't fall into the same trap because it used more computing time and obviously saw ahead.
My Sensory 9 did not get caught and gave more varied responses to not go the same route.

Dave

Sensory 9 was Spracklen vintage, so clearly lot lot stronger than the others comps. you mention, all of them by Nelson.

Playing septennial regards
Fern

Larry · Post by **Larry** » Mon Jul 20, 2015 7:40 am

Fernando wrote:
Sensory 9 was Spracklen vintage, so clearly lot lot stronger than the others comps. you mention, all of them by Nelson.

And Sensory 9 says on it's box "stronger version of...("Champion voice"),
and it's rating shows it to be around 20 elo stronger. Of course that's the
original 1.5mhz version. It also came in 1.6mhz, and eventually in an
outright different program running at 2mhz, unofficially dubbed the SC9B.
The Englishman Steve Braid made a mainboard for SC9B running at 8mhz
complete with a built in extended openings module. This one would play
at maybe elo1750 or better.
L

afos99 · Post by **afos99** » Mon Jul 20, 2015 9:01 am

But surely my point is that the voice challenger - very similar to CC7 does not fall for the trap because it uses more thinking time on level 1. I got this machine in 1980 and was always annoyed that on a particular level would think beyond that level when under pressure.

I understand the levels were quoted as an average time but that effectively means that you are playing against an opponent that could change from say level 2 to level 7 at various stages of the game. Is that unique to VCC's program and all other software will produces a move at the exact time limit?

Dave

Post by **Fernando** » Mon Jul 20, 2015 3:40 pm

afos99 wrote:But surely my point is that the voice challenger - very similar to CC7 does not fall for the trap because it uses more thinking time on level 1. I got this machine in 1980 and was always annoyed that on a particular level would think beyond that level when under pressure.

I understand the levels were quoted as an average time but that effectively means that you are playing against an opponent that could change from say level 2 to level 7 at various stages of the game. Is that unique to VCC's program and all other software will produces a move at the exact time limit?

Dave

Some comps has an option to give exact time to every move, but in tourn. level or whatever the name is, it is an average. . And as much many moves were played instantly because they were booked, that average could be lot more extensive than what you expect

Average regards
Fern