Nelson again

[Fernando]

I just saw the issue Nelson-as-an author of this or that program, coming again to light.
Let me offer an intermediate solution:
Nelson was programmer in full of old Fidelity machines and perhaps ,too, the most cheap Excalibur one.
Nelson was partially part of the program of better machines in the same way he shared the programming of Fidelity when the Spracklen were the God. If I remember well, even in those times Nelson used to appear as the chief guy in the programming side of Fidelity due to his participation in the hardware design of the machines. As much there is some programming in that too, Nelson can say he programmed this or that device just because he programmed the voice function, etc.
In intermediate machines as Alexandra and Igor I have noted that the game is similar in lot of aspects to Spracklen Champion Challenger, with some extra here and there.
The conclusion is: in these times the brains of the dedicated units are already a mix of many hands along time....

Fern

[scandien]

hello,

i was just thinking the same.. The Fidelity Champion ( which evolved later in Excellence and then Excel series) was issued from SARGONIII program.
May be the Excalibur Grand Master is related to this old Software.

Unfortunately there is no evidence of this ( and finding one could be very hard). So far I am trying to check for GrandMaster Style ( with Khmelnitsky tests) and i will run the test for Sargon 5 too, to see if the playing style are related.

Best Regards and Happy Christmas to all,

Nicolas

[Mike Watters]

I agree with you Fern. By the mid to late 1990s dedicated chess computers were mainly a corporate creation. But there is still a desire to credit the main author of the chess engine, both to understand each chess computer and to celebrate the work of those involved.

When Craig Barnes came forward and contributed on here I emailed him to get more information about who did what on SciSys/Saitek chess computers. He was due to meet Julio Kaplan and I had hopes that they would make things clearer on the division of work between them. However his later response seemed to dodge the issue. Which is why I now acknowledge both he and Kaplan jointly on my SciSys website section and leave it at that.

As for Ron Nelson, he has such an important place in the early history of chess computers that the controversy over what, if anything, he contributed to Excalibur chess programs is a pity. But typical of the internet, every time someone repeats apparently 'unjustified' claims for what he may have contributed, someone else turns up with evidence to the contrary.

On the plus side I have USSR, East German and Brazilian made chess computers all with Ron Nelson programs, which he probably never knew anything about.

[Fernando]

Hi Nicolas:

Sargon V was an entirely different, lot stronger program than Champion and even par excellence. If I am not wrong, SV was a 2300 or so Elo program, above Par and others of the Fidelity era.
Sargon V was almost the strongest the Spracklen ever produced, second only to Sparks module that was a failure due to the superiority of Lang.

Fern

[scandien]

Yes i am agree with you.. Sargon V is a strong program (on Dos Bos simulating slow hardware it is nearly 1850, not far from my evaluation for Excellence) . But Champion and Exellence were based on SARGON III, which have been probably improved to Sargon IV and V ... That's why i think they may be related ... To be check ..

[Fernando]

Yes i am agree with you.. Sargon V is a strong program (on Dos Bos simulating slow hardware it is nearly 1850, not far from my evaluation for Excellence) . But Champion and Exellence were based on SARGON III, which have been probably improved to Sargon IV and V ... That's why i think they may be related ... To be check ..

Sargon IV was a failure. Something went wrong with the code if I recall right... But not sure. I have downloaded it to play under DOS but not playing enough to test if was so or later debugged.

[Bryan Whitby]

Computer Chess Reports 1991-92 by Larry Kaufman.

Sargon V is the program that Sargon IV should have been. In other words it is pretty much a translation of the Sargon/Mac IV version, which is to say a translation of the Fidelity Mach II program. The PC version of Sargon IV was hopelessly bungled and played very poor chess relative to all serious chess playing programs (including Sargon III)

Copies of other Computer Chess Reports can be found at Mike Watters web site.
http://www.chesscomputeruk.com/html/com ... ports.html

[Fernando]

Computer Chess Reports 1991-92 by Larry Kaufman.

Sargon V is the program that Sargon IV should have been. In other words it is pretty much a translation of the Sargon/Mac IV version, which is to say a translation of the Fidelity Mach II program. The PC version of Sargon IV was hopelessly bungled and played very poor chess relative to all serious chess playing programs (including Sargon III)

Copies of other Computer Chess Reports can be found at Mike Watters web site.
http://www.chesscomputeruk.com/html/com ... ports.html

Yes, I remember that lines by Kaufman. From curiosity it was i got sargon Iv to see howmuch "bungled" the thing was....

[scandien]

I have just finish to apply the test to Sargon and to GrandMaster.

I was wrong .. the program are different . The grandmaster doesn't solve problem easily solved by the SARGON!
Their style is very different, as if the SARGON is ok in endgame, calcul, and strategy and this is not the case for the GrandMaster .

The GrandMaster strength rely on Threat detection and tactic and is really too weak in Strategy, Endgame and Attack !

Best regards

Nicolas


reminder :
Endgame : A phase where player often don't take care ( if we compare with the opening or middle game) and this is the same for the machines. Mastering the endgames requires teorical knowledges, strategic capability, tactl, and good capabilities for calculation.

Middle Game : This is the more decisive part of the game. Nearly 70% of the game are solved during this phase. This phase requires tactical and strategical capability. Generally humain players follow basic plan, but quicly the players have to decide if they have to attack or to concentrate in defense.



Calculation: Mainly this is the skill to anticipate opponent's moves.

Strategy: During a game, when we play a move we should have a good reason to play it ! Developing a plan enable the player to find the good move in positions without combinative motive.
The player have to enhance his skills in evaluation of a position, elabooration of short term plan ( and after long term planification can be mastered) , reveal opponent's plan , and finally find candidate moves.

Tactic : Some tremandous move can lead to forced variation, generally short. At the source of those move , there are combinatives motives.To master the tactic in chess , the player have to know as many motive as possible to be able to gain benefit for all efforts previously done during the game.

threat : Threats identification enable a player to take counter opponent's threat, and to create his own attack. Threats can be of different sort, straegical or tactical, and can occur during any phase of the game.

Attack : This is mainly the ability to identify a target, and to prepare his force to reach it. This is not a threat , as this is mainly a long term operation, where the force have to be regroup. This is basically a strategic skill.


Counter-Attack: The capability , where we are under attack , to change the course of the game and to launch his own attack! A bad couter attack generally leads to severe defeat!.

Defense: In bad position a player should be able to hold, and only after to found a way to more agressive path. The Defense should not be passive and must stay dynamical ( to be able to launch a counter-attack) . This skill enable the plyer to save very difficult position.

[Fernando]

I have just finish to apply the test to Sargon and to GrandMaster.

I was wrong .. the program are different . The grandmaster doesn't solve problem easily solved by the SARGON!
Their style is very different, as if the SARGON is ok in endgame, calcul, and strategy and this is not the case for the GrandMaster .

The GrandMaster strength rely on Threat detection and tactic and is really too weak in Strategy, Endgame and Attack !

Best regards

Nicolas

Not amazed. GM by Excalibur -I have one- is way under sargon. It plays not that bad, but lacks sharpness and acumen. It is clearly a Nelson thing with some extra given by some other code and/or the common programming knowledge that every one can get these days just looking any engine inside.
Spracklen stuff was always sharp and interesting.

F

reminder :
Endgame : A phase where player often don't take care ( if we compare with the opening or middle game) and this is the same for the machines. Mastering the endgames requires teorical knowledges, strategic capability, tactl, and good capabilities for calculation.

Middle Game : This is the more decisive part of the game. Nearly 70% of the game are solved during this phase. This phase requires tactical and strategical capability. Generally humain players follow basic plan, but quicly the players have to decide if they have to attack or to concentrate in defense.



Calculation: Mainly this is the skill to anticipate opponent's moves.

Strategy: During a game, when we play a move we should have a good reason to play it ! Developing a plan enable the player to find the good move in positions without combinative motive.
The player have to enhance his skills in evaluation of a position, elabooration of short term plan ( and after long term planification can be mastered) , reveal opponent's plan , and finally find candidate moves.

Tactic : Some tremandous move can lead to forced variation, generally short. At the source of those move , there are combinatives motives.To master the tactic in chess , the player have to know as many motive as possible to be able to gain benefit for all efforts previously done during the game.

threat : Threats identification enable a player to take counter opponent's threat, and to create his own attack. Threats can be of different sort, straegical or tactical, and can occur during any phase of the game.

Attack : This is mainly the ability to identify a target, and to prepare his force to reach it. This is not a threat , as this is mainly a long term operation, where the force have to be regroup. This is basically a strategic skill.


Counter-Attack: The capability , where we are under attack , to change the course of the game and to launch his own attack! A bad couter attack generally leads to severe defeat!.

Defense: In bad position a player should be able to hold, and only after to found a way to more agressive path. The Defense should not be passive and must stay dynamical ( to be able to launch a counter-attack) . This skill enable the plyer to save very difficult position.

[Monsieur Plastique]

In intermediate machines as Alexandra and Igor I have noted that the game is similar in lot of aspects to Spracklen Champion Challenger, with some extra here and there.

You've hit the nail on the head here Fern. These have always been my observations of these later "Nelson" Excaliburs. I might have mentioned some time back that although I still believe they are essentially his "work", there is a suspicion deep down to my mind that he developed everything off some very old early 1980s code that perhaps the Spraklens discarded and in one way or another let Nelson run with it as he pleased. After all, why would the Spraklens have wanted anything to do with their Champion Challenger code when they were starting to deliver next generation and far better products such as the Sensory 9 and soon after the excellent...um..Excellence.

Here is a game I played against Alexandra on it's most difficult tournament level. If I did not tell you it was Alexandra it could for all the world have been a Champion Sensory Challenger:

[Event "40 moves in 2 hours"]
[Site "?"]
[Date "2013.08.10"]
[Round "?"]
[White "Parle, Jonathan"]
[Black "Excalibur Alexandra"]
[Result "1-0"]
[ECO "D45"]
[PlyCount "65"]

1. d4 d5 2. c4 c6 3. Nc3 Nf6 4. e3 e6 5. Nf3 Nbd7 6. Be2 Bd6 7. cxd5 cxd5 8.
O-O O-O 9. a3 Kh8 10. b4 a5 11. b5 Nb6 12. Bb2 Nc4 13. Bxc4 dxc4 14. Qe2 Qc7
15. e4 Nd7 16. e5 Be7 17. d5 Bd8 18. d6 Qb8 19. Nd4 Re8 20. f4 f6 21. Qh5 g6
22. Qh6 Qa7 23. Nce2 Kg8 24. Kh1 fxe5 25. fxe5 Nxe5 26. Nxe6 Bxe6 27. b6 c3 28.
Bxc3 Qa6 29. Bxe5 Bf6 30. Rxf6 Re7 31. dxe7 Bf7 32. Rxg6+ Bxg6 33. Qg7# 1-0

[Fernando]

Well, yes, Champion was not the great thing, BUT I LOVE IT a lot because it was my first chess machine that played a real game of chess instead of the preposterous chess played by precedent machines.... In even today if you play it carelessly, you can suffer a nasty surprise....

Fern

[scandien]

Hello Mister Plastic,

i saw in ICGA homepage that the SC9 series were based on SARGON III program, and The EXCEl MACH II is based on SARGON V ( the latest and best Spraklens program).

So the Alexandra ( very weak machine from my point of view) is related to excel ???? ...

I don't really understand the rating of the SARGON V in my list, because in problem resolutuon it seems really strong .. probably this is because i have only a small opening book.

Best regards

Nicolas

[IanO]

So the Alexandra ( very weak machine from my point of view) is related to excel ???? ...

No, the license for the Spracklen engines went to Mephisto when they bought Fidelity. The Excalibur engines are all of questionable pedigree (possibly Horvath, possibly Nelson), but certainly not of any of the major authors like the Spracklens.

Different chips, too. Spracklen's stuff ran on 6502, then 68000. Excalibur's computers are all H8. Completely different architectures.

[Monsieur Plastique]

So the Alexandra ( very weak machine from my point of view) is related to excel ???? ...

No, if it were related to anything it would be the original Fidelity Champion Sensory Challenger - the precursor to the Sensory 9 and a program that even in the day was criticised for it's unhumanlike and poor positional play.

No one has done a serious and robust rating test of Alexandra on tournament level (i.e at least 40 x 40 in 2 games) but in my estimation it is probably around the mid 1500s Swedish ELO. It's probably around 30 points weaker than a Sensory 9 and definitely not up to the standard of the original 2 MHz Constellation.

And funnily enough, if you took that original Fidelity Champion Sensory and simply upgraded it to play on a 10 Mhz H8 (as is the case with Alexandra), I reckon it would be around the mid to high 1500s Swedish as well!