Novag Obsidian vs Saitek Travel Champion 2100

Designed for posting all types of tournaments and Games (e.g. Man vs. Machine, Computer vs. Computer and basement matches.)

Moderators: Harvey Williamson, Watchman

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
Post Reply
User avatar
sje
Full Member
Posts: 639
Joined: Sat Sep 28, 2013 2:28 am
Location: Land of Snow, Mud, and Bugs, NH USA

Post by sje »

One should remember that most dedicated chess computers like the Novag Obsidian have been optimized for rapid games, not games with very long time controls. There may be some exceptions, but none among machines with less than 1 KB RAM.

--------

If White plays 42 Rc4, then Symbolic predicts:
[+5.525/26/1:13:01.485/5,364,113,842/6,223] 42... Rexg2+ 43 Nxg2 Rxg2+ 44 Kh1 Re2+ 45 Kg1 Re1+ 46 Kf2 Rxc1 47 Ke3 Be4 48 Rc8+ Kg7 49 Rc7+ Kf8 50 Kxe4 Re1+ 51 Kd5 c1=Q 52 Rxc1 Rxc1 53 h4 Kg7 54 h5 Rf1 55 Ke6 h6 56 Kd7 Rxf5 57 Kc6 Rxh5

Node frequency: 1.22 MHz
RadioSmall
Member
Posts: 273
Joined: Fri Jul 02, 2010 7:12 pm

Post by RadioSmall »

sje wrote:One should remember that most dedicated chess computers like the Novag Obsidian have been optimized for rapid games, not games with very long time controls. There may be some exceptions, but none among machines with less than 1 KB RAM.

--------

If White plays 42 Rc4, then Symbolic predicts:
[+5.525/26/1:13:01.485/5,364,113,842/6,223] 42... Rexg2+ 43 Nxg2 Rxg2+ 44 Kh1 Re2+ 45 Kg1 Re1+ 46 Kf2 Rxc1 47 Ke3 Be4 48 Rc8+ Kg7 49 Rc7+ Kf8 50 Kxe4 Re1+ 51 Kd5 c1=Q 52 Rxc1 Rxc1 53 h4 Kg7 54 h5 Rf1 55 Ke6 h6 56 Kd7 Rxf5 57 Kc6 Rxh5

Node frequency: 1.22 MHz
The Novag Obsidian has this bug of playing a move way too early ( in fact at ply 11 ) when set to infinite time .... Also interestingly another chess computer I own ( The RadioShack Chessmaster ) has got the same problem... Although I cannot establish when or why this happens ( because it does not happen in every move ........)... A good question to ask about dedicated Comps in general is this :how much Elo does one gain going from say 40 moves in 2 hours to 24 hours per move??
User avatar
sje
Full Member
Posts: 639
Joined: Sat Sep 28, 2013 2:28 am
Location: Land of Snow, Mud, and Bugs, NH USA

Post by sje »

RadioSmall wrote:The Novag Obsidian has this bug of playing a move way too early ( in fact at ply 11 ) when set to infinite time
It's not a bug, but rather a sign of running low on memory due to a complex position.
RadioSmall
Member
Posts: 273
Joined: Fri Jul 02, 2010 7:12 pm

Post by RadioSmall »

sje wrote:
RadioSmall wrote:The Novag Obsidian has this bug of playing a move way too early ( in fact at ply 11 ) when set to infinite time
It's not a bug, but rather a sign of running low on memory due to a complex position.
But I do not have this problem with many other Chess Computers I own for example the RadioShack 2250 XL ...... (which has never exhibited this behavoir in years of 24 hours per move testing.......)
User avatar
sje
Full Member
Posts: 639
Joined: Sat Sep 28, 2013 2:28 am
Location: Land of Snow, Mud, and Bugs, NH USA

Post by sje »

RadioSmall wrote:But I do not have this problem with many other Chess Computers I own for example the RadioShack 2250 XL
That machine may have a different algorithm which does not terminate when no further information can be had by continuing the search.
RadioSmall
Member
Posts: 273
Joined: Fri Jul 02, 2010 7:12 pm

Post by RadioSmall »

sje wrote:
RadioSmall wrote:But I do not have this problem with many other Chess Computers I own for example the RadioShack 2250 XL
That machine may have a different algorithm which does not terminate when no further information can be had by continuing the search.
Also please note that the Designer Master 2265 is also free from this problem....
User avatar
sje
Full Member
Posts: 639
Joined: Sat Sep 28, 2013 2:28 am
Location: Land of Snow, Mud, and Bugs, NH USA

Post by sje »

sje wrote:If White plays 42 Rc4, then Symbolic predicts:
[+5.525/26/1:13:01.485/5,364,113,842/6,223] 42... Rexg2+ 43 Nxg2 Rxg2+ 44 Kh1 Re2+ 45 Kg1 Re1+ 46 Kf2 Rxc1 47 Ke3 Be4 48 Rc8+ Kg7 49 Rc7+ Kf8 50 Kxe4 Re1+ 51 Kd5 c1=Q 52 Rxc1 Rxc1 53 h4 Kg7 54 h5 Rf1 55 Ke6 h6 56 Kd7 Rxf5 57 Kc6 Rxh5
[+5.792/28/10:09:31.052/46,539,978,358/67,013] 42... Rexg2+ 43 Nxg2 Rxg2+ 44 Kh1 Re2+ 45 Kg1 Re1+ 46 Kf2 Rxc1 47 Ke2 Be4 48 Rxe4 Rd1 49 Re8+ Kg7 50 Rc8 c1=Q 51 Rxc1 Rxc1 52 Kd3 Rh1 53 Kc4 Rh2 54 a4 a5 55 b4 axb4 56 Kxb4 Rxh3 57 a5 Rf3 58 a6 Rxf5

Node frequency: 1.27 MHz
Node period: 786 ns (2.09 K CPU cycles)
RadioSmall
Member
Posts: 273
Joined: Fri Jul 02, 2010 7:12 pm

Post by RadioSmall »

sje wrote:
RadioSmall wrote:But I do not have this problem with many other Chess Computers I own for example the RadioShack 2250 XL
That machine may have a different algorithm which does not terminate when no further information can be had by continuing the search.
You are indeed very correct about this..............
Wardy
Full Member
Posts: 940
Joined: Tue Jul 31, 2007 7:17 pm
Location: Wellingborough

Post by Wardy »

Having just beaten the citrine for the first time it would appear the resignation function doesn't work. Even at mate in one it carried on playing.

Looks like you'll have to make the call manually, plus thinking about it logically would the machine resign in infinite mode anyway?
User avatar
sje
Full Member
Posts: 639
Joined: Sat Sep 28, 2013 2:28 am
Location: Land of Snow, Mud, and Bugs, NH USA

Post by sje »

[Event "Correspondence Game"]
[Site "http://hiarcs.net/forums/viewtopic.php?p=62613"]
[Date "2013.10.30"]
[Round "-"]
[White "Novag Obsidian"]
[Black "Saitek Travel Champion 2100"]
[Result "*"]

1 e4 e5 2 Nf3 Nc6 3 Bb5 a6 4 Ba4 Nf6 5 O-O Be7 6 Re1 b5 7 Bb3 O-O 8 c3 d6 9 h3
Na5 10 Bc2 c5 11 d4 cxd4 12 cxd4 Bb7 13 Nc3 Qc7 14 dxe5 dxe5 15 Nd5 Nxd5 16
exd5 Rad8 17 Rxe5 Nc4 18 Qd3 g6 19 Re1 Bf6 20 Bb3 Bxd5 21 Bxc4 bxc4 22 Qe2 c3
23 b3 c2 24 Ne5 Rfe8 25 f4 Bh4 26 Rf1 f6 27 Qg4 Qc5+ 28 Kh2 Qc3 29 Nxg6 Qxa1 30
Nxh4+ Kh8 31 Qh5 Be4 32 f5 Qe5+ 33 Rf4 Qd6 34 Qh6 Rd7 35 Qh5 Red8 36 Qe2 Re7 37
Qh5 Rg7 38 Qe2 Re8 39 Qc4 Bb7 40 Qd4 Qxd4 41 Rxd4 Re2 42 Rd8+ *

[fen]3R3k/1b4rp/p4p2/5P2/7N/1P5P/P1p1r1PK/2B5 b - - 2 42[/fen]
User avatar
sje
Full Member
Posts: 639
Joined: Sat Sep 28, 2013 2:28 am
Location: Land of Snow, Mud, and Bugs, NH USA

Post by sje »

Wardy wrote:Having just beaten the citrine for the first time it would appear the resignation function doesn't work. Even at mate in one it carried on playing.

Looks like you'll have to make the call manually, plus thinking about it logically would the machine resign in infinite mode anyway?
My Citrine has resigned hundreds of times in games played over its serial link against Symbolic.

The Obsidian really doesn't have an infinite mode, just a 16 ply depth mode. Symbolic doesn't have a fixed depth limit and could go a million ply deep. Note that the longest possible game is less than 13,000 ply.
Wardy
Full Member
Posts: 940
Joined: Tue Jul 31, 2007 7:17 pm
Location: Wellingborough

Post by Wardy »

Search started.

Perhaps I missed the resignation on the screen in my excitment but I did think I was checking.
User avatar
sje
Full Member
Posts: 639
Joined: Sat Sep 28, 2013 2:28 am
Location: Land of Snow, Mud, and Bugs, NH USA

Post by sje »

When using the serial link, the Citrine reports a game over code which is a prefix followed by one of several integer values. One of these means "I resign" and that's what my program saw. A resignation comes fairly late.
User avatar
sje
Full Member
Posts: 639
Joined: Sat Sep 28, 2013 2:28 am
Location: Land of Snow, Mud, and Bugs, NH USA

Post by sje »

Also, a resignation comes only once per game. I believe the same is true of non stalemate draws.

There are six different game over codes emitted by the Citrine:

Draws (four kinds)
Checkmate
Resignation

Because of its RAM size limit, I doubt if the Citrine cannot always correctly claim position repetition. It would have to store hash codes for positions from the last hundred ply and there just isn't room for this.
Wardy
Full Member
Posts: 940
Joined: Tue Jul 31, 2007 7:17 pm
Location: Wellingborough

Post by Wardy »

The TC plays Rg8 with a score of 3.3 It managed to complete a 13 ply search but didn't make much of a dent at 14 ply. The small number of candidate moves obviously helping to reduce the size of the search tree.

[fen]3R2rk/1b5p/p4p2/5P2/7N/1P5P/P1p1r1PK/2B5 w - - 0 43[/fen]

1. e4 e5 2. Nf3 Nc6 3. Bb5 a6 4. Ba4 Nf6 5. O-O Be7 6. Re1 b5 7. Bb3 O-O 8. c3
d6 9. h3 Na5 10. Bc2 c5 11. d4 cxd4 12. cxd4 Bb7 13. Nc3 Qc7 14. dxe5 dxe5 15.
Nd5 Nxd5 16. exd5 Rad8 17. Rxe5 Nc4 18. Qd3 g6 19. Re1 Bf6 20. Bb3 Bxd5 21.
Bxc4 bxc4 22. Qe2 c3 23. b3 c2 24. Ne5 Rfe8 25. f4 Bh4 26. Rf1 f6 27. Qg4 Qc5+
28. Kh2 Qc3 29. Nxg6 Qxa1 30. Nxh4+ Kh8 31. Qh5 Be4 32. f5 Qe5+ 33. Rf4 Qd6 34.
Qh6 Rd7 35. Qh5 Red8 36. Qe2 Re7 37. Qh5 Rg7 38. Qe2 Re8 39. Qc4 Bb7 40. Qd4
Qxd4 41. Rxd4 Re2 42. Rd8+ Rg8 *
Post Reply