Sargon does NOT know the opposition principle, you can view it here and replay what Sargon played.
Sargon 1978 UCI Available
Moderators: hgm, Dann Corbit, Harvey Williamson
Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
-
Chessqueen
- Posts: 1332
- Joined: Wed Sep 05, 2018 12:16 am
- Full name: J.L.P.Q
-
Chessqueen
- Posts: 1332
- Joined: Wed Sep 05, 2018 12:16 am
- Full name: J.L.P.Q
Re: Sargon 1978 UCI Available
Here is another game where Sargon was winning versus Cicada rated 1538 but, once more repeated the three fold repetition bug, I will continue this game using Rainman to see how quickly Rainman rated 1619 finish off Cicadarcmaddox wrote: ↑Wed May 27, 2020 2:58 amBill Forster, author of the Tarrasch GUI, has released a UCI version of Sargon, the well known program from the Spracklens.
I had a copy of Sargon in the early 80's running on my Commodore VIC-20, using a black and white television for a monitor! It was my first chess software that wasn't in a dedicated chess computer.
So here ya go:
Main Tarrasch GUI Website ---->> http://www.triplehappy.com/
Link to Releveant Blog Post ---->> https://triplehappy.wordpress.com/
-
Chessqueen
- Posts: 1332
- Joined: Wed Sep 05, 2018 12:16 am
- Full name: J.L.P.Q
Re: Sargon 1978 UCI Available
Chessqueen wrote: ↑Sun Jan 10, 2021 5:48 pmHere is another game where Sargon was winning versus Cicada rated 1538 but, once more repeated the three fold repetition bug, I will continue this game using Rainman to see how quickly Rainman rated 1619 finish off Cicadarcmaddox wrote: ↑Wed May 27, 2020 2:58 amBill Forster, author of the Tarrasch GUI, has released a UCI version of Sargon, the well known program from the Spracklens.
I had a copy of Sargon in the early 80's running on my Commodore VIC-20, using a black and white television for a monitor! It was my first chess software that wasn't in a dedicated chess computer.
So here ya go:
Main Tarrasch GUI Website ---->> http://www.triplehappy.com/
Link to Releveant Blog Post ---->> https://triplehappy.wordpress.com/
I decided to play it out myself and I took the position and play with Black pieces
-
Bill Forster
- Posts: 42
- Joined: Mon Sep 21, 2015 5:47 am
- Location: New Zealand
- Contact:
Re: Sargon 1978 UCI Available
You are quite correct, Sargon does not know the opposition principle. Sargon doesn't know anything really, it is just a thorough (but hence inefficient and bad) calculator that knows nothing except the basic rules. As you have found, it doesn't even know the less basic rules like three-fold repetition. But Sargon is not quite so bad that it can't CALCULATE the opposition principle, if you give it enough time. From the project page https://github.com/billforsternz/retro-sargon;Chessqueen wrote: ↑Sat Jan 09, 2021 8:37 pmSargon does NOT know the opposition principle, you can view it here and replay what Sargon played.
"It might sound that extending Sargon's search depth well beyond 6 hasn't been very useful because the exponential growth makes levels beyond 8 or so inaccessible in practice. This would be true if chess stopped in the middlegame, but as more and more pieces come off the board, there are less moves to look at in each position, and Sargon can see deeper in a reasonable time. This is the real beauty of adaptive rather than fixed depth. As a simple example consider the position with White (to play) king on e4, pawn on d4, Black king on d7. A serious human player will instantaneously play Kd5 here, knowing that this is the only way to eventually force pawn promotion. Sargon, with absolutely no chess knowledge, just the ability to calculate variations in search of extra material or board control, plays Kd5 at depth 13 (or more). The PV associated with this move features the pawn queening triumphantly right at the end of the line. It takes 3 minutes and 30 seconds to make the calculation on my computer. It's hardly world shattering but it's vastly beyond the capacity of the original Z80 implementation, and shows what a faster CPU and more memory can do to an otherwise fixed chess calculation algorithm."
You should recognise the position described and the move played as being a basic illustration of seizing the opposition.
Well I gave myself "at least a week" on December 31st. I was careful not to make promises. I have actually implemented the fix, as anticipated it was not an easy job. I haven't released it yet because I want to do more testing and polishing. In the meantime I have taken some time out to play some chess in the New Zealand championships which is delaying things. I hope you don't mind. You shouldn't expect too much. Sargon will no longer blindly walk into repetitions unnecessarily, but it hasn't become any more skilled at converting winning positions. It needs either a more sophisticated scoring function or greater calculation depth so that it can see the benefit of penetrating with the king or whatever longer term plan will eventually break the opponent.Chessqueen wrote: ↑Sat Jan 09, 2021 7:03 amBill how much longer will it take you to correct the repetition bug in Sargon so I can play another 10 games, since Sargon is so close to my rating ?
In general I appreciate your interest in my project, but I wish you would actually read about what I have done and why (hint: giving you an ideal training partner was not one of my goals
). For example;Given permission? Did the Spracklens join the Talkchess discussion? NNUE????!?. Actually although NNUE grafted onto Sargon is totally insane, who knows it could even be possible.Chessqueen wrote: ↑Sun Jan 03, 2021 4:33 amSince you have been given permission to work or fix Sargon Try to give it NN or NNU,E and fix the evaluation by giving it an extra 100+ ratings points
-
mclane
- Posts: 18368
- Joined: Thu Mar 09, 2006 5:40 pm
- Location: US of Europe, germany
- Full name: Thorsten Czub
- Contact:
Re: Sargon 1978 UCI Available
We should contact them. They enhanced the engine over many years.
The strongest was maybe the sparc module in the Saitek Renaissance.
The strongest was maybe the sparc module in the Saitek Renaissance.
What seems like a fairy tale today may be reality tomorrow.
Here we have a fairy tale of the day after tomorrow....
Here we have a fairy tale of the day after tomorrow....
-
Chessqueen
- Posts: 1332
- Joined: Wed Sep 05, 2018 12:16 am
- Full name: J.L.P.Q
Re: Sargon 1978 UCI Available
If you have NOT released it yet. it has to be because you are making other changes, After playing 10 games the score was 5 win 2 draw and 3 lost in my favor, but the two draws games Sargon was clearly winning against me, and decided to repeat three times, which NOW makes Sargon a solid close to 1900 or possibly better on a good computer like inter i9 9900K or i9 10900K but of course Sargon can only uses 1 core, NOT multiple.Bill Forster wrote: ↑Mon Jan 11, 2021 7:06 pmYou are quite correct, Sargon does not know the opposition principle. Sargon doesn't know anything really, it is just a thorough (but hence inefficient and bad) calculator that knows nothing except the basic rules. As you have found, it doesn't even know the less basic rules like three-fold repetition. But Sargon is not quite so bad that it can't CALCULATE the opposition principle, if you give it enough time. From the project page https://github.com/billforsternz/retro-sargon;Chessqueen wrote: ↑Sat Jan 09, 2021 8:37 pmSargon does NOT know the opposition principle, you can view it here and replay what Sargon played.
"It might sound that extending Sargon's search depth well beyond 6 hasn't been very useful because the exponential growth makes levels beyond 8 or so inaccessible in practice. This would be true if chess stopped in the middlegame, but as more and more pieces come off the board, there are less moves to look at in each position, and Sargon can see deeper in a reasonable time. This is the real beauty of adaptive rather than fixed depth. As a simple example consider the position with White (to play) king on e4, pawn on d4, Black king on d7. A serious human player will instantaneously play Kd5 here, knowing that this is the only way to eventually force pawn promotion. Sargon, with absolutely no chess knowledge, just the ability to calculate variations in search of extra material or board control, plays Kd5 at depth 13 (or more). The PV associated with this move features the pawn queening triumphantly right at the end of the line. It takes 3 minutes and 30 seconds to make the calculation on my computer. It's hardly world shattering but it's vastly beyond the capacity of the original Z80 implementation, and shows what a faster CPU and more memory can do to an otherwise fixed chess calculation algorithm."
You should recognise the position described and the move played as being a basic illustration of seizing the opposition.
Well I gave myself "at least a week" on December 31st. I was careful not to make promises. I have actually implemented the fix, as anticipated it was not an easy job. I haven't released it yet because I want to do more testing and polishing. In the meantime I have taken some time out to play some chess in the New Zealand championships which is delaying things. I hope you don't mind. You shouldn't expect too much. Sargon will no longer blindly walk into repetitions unnecessarily, but it hasn't become any more skilled at converting winning positions. It needs either a more sophisticated scoring function or greater calculation depth so that it can see the benefit of penetrating with the king or whatever longer term plan will eventually break the opponent.Chessqueen wrote: ↑Sat Jan 09, 2021 7:03 amBill how much longer will it take you to correct the repetition bug in Sargon so I can play another 10 games, since Sargon is so close to my rating ?
In general I appreciate your interest in my project, but I wish you would actually read about what I have done and why (hint: giving you an ideal training partner was not one of my goals
Given permission? Did the Spracklens join the Talkchess discussion? NNUE????!?. Actually although NNUE grafted onto Sargon is totally insane, who knows it could even be possible.Chessqueen wrote: ↑Sun Jan 03, 2021 4:33 amSince you have been given permission to work or fix Sargon Try to give it NN or NNU,E and fix the evaluation by giving it an extra 100+ ratings points
NOTE: Can you make it part of the engines that comes with Tarrasch GUI
https://www.triplehappy.com/-
Chessqueen
- Posts: 1332
- Joined: Wed Sep 05, 2018 12:16 am
- Full name: J.L.P.Q
Re: Sargon 1978 UCI Available
I just finished another 20 games Against Sargon the unfix one with repetition bug game time control 15'+5" and the final score of these 20 games
14 win 3 lost and 3 draw in my favor. Some people enjoy playing handicap versus Komodo, but I found out that playing versus engines near your rating is more fun even if they are primitive engine from the late 78 . I am also playing versus Grizzly 1.40 1.b rated 1786 and so far the score is 4 win 2 lost and 2 draw in my favor
-
Chessqueen
- Posts: 1332
- Joined: Wed Sep 05, 2018 12:16 am
- Full name: J.L.P.Q
Re: Sargon 1978 UCI Available
Bill Forster wrote: ↑Mon Jan 11, 2021 7:06 pmYou are quite correct, Sargon does not know the opposition principle. Sargon doesn't know anything really, it is just a thorough (but hence inefficient and bad) calculator that knows nothing except the basic rules. As you have found, it doesn't even know the less basic rules like three-fold repetition. But Sargon is not quite so bad that it can't CALCULATE the opposition principle, if you give it enough time. From the project page https://github.com/billforsternz/retro-sargon;Chessqueen wrote: ↑Sat Jan 09, 2021 8:37 pmSargon does NOT know the opposition principle, you can view it here and replay what Sargon played.
"It might sound that extending Sargon's search depth well beyond 6 hasn't been very useful because the exponential growth makes levels beyond 8 or so inaccessible in practice. This would be true if chess stopped in the middlegame, but as more and more pieces come off the board, there are less moves to look at in each position, and Sargon can see deeper in a reasonable time. This is the real beauty of adaptive rather than fixed depth. As a simple example consider the position with White (to play) king on e4, pawn on d4, Black king on d7. A serious human player will instantaneously play Kd5 here, knowing that this is the only way to eventually force pawn promotion. Sargon, with absolutely no chess knowledge, just the ability to calculate variations in search of extra material or board control, plays Kd5 at depth 13 (or more). The PV associated with this move features the pawn queening triumphantly right at the end of the line. It takes 3 minutes and 30 seconds to make the calculation on my computer. It's hardly world shattering but it's vastly beyond the capacity of the original Z80 implementation, and shows what a faster CPU and more memory can do to an otherwise fixed chess calculation algorithm."
You should recognise the position described and the move played as being a basic illustration of seizing the opposition.
Well I gave myself "at least a week" on December 31st. I was careful not to make promises. I have actually implemented the fix, as anticipated it was not an easy job. I haven't released it yet because I want to do more testing and polishing. In the meantime I have taken some time out to play some chess in the New Zealand championships which is delaying things. I hope you don't mind. You shouldn't expect too much. Sargon will no longer blindly walk into repetitions unnecessarily, but it hasn't become any more skilled at converting winning positions. It needs either a more sophisticated scoring function or greater calculation depth so that it can see the benefit of penetrating with the king or whatever longer term plan will eventually break the opponent.Chessqueen wrote: ↑Sat Jan 09, 2021 7:03 amBill how much longer will it take you to correct the repetition bug in Sargon so I can play another 10 games, since Sargon is so close to my rating ?
In general I appreciate your interest in my project, but I wish you would actually read about what I have done and why (hint: giving you an ideal training partner was not one of my goals
Given permission? Did the Spracklens join the Talkchess discussion? NNUE????!?. Actually although NNUE grafted onto Sargon is totally insane, who knows it could even be possible.Chessqueen wrote: ↑Sun Jan 03, 2021 4:33 amSince you have been given permission to work or fix Sargon Try to give it NN or NNU,E and fix the evaluation by giving it an extra 100+ ratings points
This is probably the most interesting game among lower end engine below 2000 Elo. The first 6 move I decided to play it for both engines to study this Villemson Gambit Opening, I did NOT wanted to use a top engine Like Ethereal, or Komodo 13.3, since I wanted to see what engine close to my rating specially Heracles did against Sargon at 6 depth, since Sargon at 8 depth does NOT work good and it takes too much time sometimes 3 minutes to make a single move. Well probably against an engine rated around 1780 to 1790 Sargon can save a draw with this Opening Villemson Gambit. My trainer always say to us students, if you are going to get into any Gambit you better knows all variations and at least calculate up to 10 moves ahead in the middlegame to equalize
-
Chessqueen
- Posts: 1332
- Joined: Wed Sep 05, 2018 12:16 am
- Full name: J.L.P.Q
Re: Sargon 1978 UCI Available
I noticed that Sargon was Unable to hold at least a draw, So i decided to play against Heracles but without endgame TB to see if I could beat it, I came close but all that I accomplished was a draw, and I do NOT know how to convert a win with a Queen versus a Rook, which if I remembered correctly as long as the Rook is close to the King it can hold an easy draw Vs my Queen.