Sunday, September 16, 2007
Vortex vs. SMIRF Blog Game #1
This area was where Gothic Vortex played the SMIRF Engine in a long time control game where each engine had up to 24 hours to make a move. Most moves were made much more quickly than this. SMIRF won with a checkmate in 73 moves.
You can replay the game one move at a time by clicking this link:
Vortex vs. SMIRF
[Event "One Move Per Day: Blog Game"]
[Site "GothicChess.BlogSpot.com"]
[Date "2007.10.07"]
[Time "00:25:48"]
[Round "Blog-Game-1"]
[White "Gothic Vortex"]
[Black "Smirf MS-168i"]
[Result "0-1"]
[Annotator "RS"]
[SetUp "1"]
[FEN "rnbqckabnr/pppppppppp/10/10/10/10/PPPPPPPPPP/RNBQCKABNR w KQkq - 0 1"]
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3 Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5 Cxd5 15. Qxd5 cxd5 16. Af3 e6 17. O-O b6 18. Ah4 j5 19. Ni5 hxi5 20. Axi5+ Ke7 21. Axh7 Rj6 22. h4 Nc5 23. h5 g5 24. h6 Ba6 25. Ag6+ Kd6 26. e3 Be2 27. a4 Bf3 28. Rh2 Be4 29. Ah5 Bxc2 30. a5 Rb8 31. axb6 axb6 32. f4 Bh7 33. fxg5 fxg5 34. Af7+ Kc6 35. Rc1 g4 36. Ag5 Rj7 37. Af4 Rg8 38. Ah5 Rj8 39. Af7 Ra8 40. Rd2 j4 41. Ag5 Be4 42. Rg1 j3 43. i3 Be5 44. Ai4 b5 45. Ah5 b4 46. Axg4 d6 47. Rdd1 b3 48. Ah5 b2 49. i4 Rji8 50. Kh2 Rah8 51. Af7 Bf6 52. Axh8 Rxh8 53. i5 Ri8 54. Ri1 Bxj2 55. Rj1 Bg5 56. Rxj3 Rxi5 57. g4 Ri6 58. Rjj1 Rxh6+ 59. Ki2 Bf3 60. Rb1 Bxg4+ 61. Ki1 Ri6+ 62. Kh1 Ri2 63. Rj4 Bf5 64. Rb4 Ne4 65. Rxe4 Re2 66. Rb4 Bxb1 67. e4 Bxe4+ 68. Kg1 b1=Q+ 69. Rxb1 Be3+ 70. Kf1 Rf2+ 71. Ke1 Bxb1 72. Kd1 Bd3 73. Ke1 Rf1# 0-1
Subscribe to:
Post Comments (Atom)
273 comments:
1 – 200 of 273 Newer› Newest»1. Nh3
1...g6 (no opening library yet)
2. d4 (book move)
Oh this is gonna be good! I got the page bookmarked.
1. Nh3, g6
2. d4, Nh6
Finally some action on here, kewl!
Is there a Vegas pool where we can bet on this? :)
3. g3
Where the heck have all you people been all this weeks?
1. Nh3, g6
2. d4, Nh6
3. g3, c6
Me, I been playing with my new gurlfriend all summer. Who has time for these silly boi games ;)
I wish Vortex would hurry up and move already.
:D
Are we allowed to try and guess the next move like they do on ICC?
4. Bf3
Sorry no diagram this time, had to rush. And yes, if people want to predict moves, that is ok.
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6
Go smurfy, beat Vortex, go smurfy, go smurfy
:)
Nice to see everyone back again. Is this a new Vortex version of a new smirf version? What machines are they being played on?
Hello Cartaphilus, it is indeed a newer SMIRF version, though I did not make a lot with it during the last time, because I first intended to stop the SMIRF project completly. Now I have heard, that its new setup would even work under MS Vista. SMIRF has a unique engine playing all variants offered in its GUI. Thus there is no special Gothic Chess version of SMIRF. It simply is a internal version, where its GC ability has been opened. Additionally here is playing the SMIRF bonus version (normally intended as a bonus for project donators), which is about 50% faster than the donationware version and has no starting nag screen. I am actually playing on a Windows XP system upon an Apple iMac BootCamp partition using an Intel T7200 @ 2GHz, but SMIRF still is a singlecore release.
5. Bf4
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6
6. Be3
The "Bishop lunge" problem in Vortex, wanting to push its Bishops too far in the opening, now it retreats.
Vortex did right so, much better than to wait for g5 and g4! SMIRF now has been believing to have won the initiative in the opening stage. - As usual SMIRF did its calculations for about one hour. It decided then to prefer Qb6 instead of d5, which had been predicted by Vortex after nine hours. I am afraid of the black Queen to be exposed too early, but SMIRF did not hesitate with that ...
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6
smirf won the opening initiative? Are we both looking at the same diagram? I see black having 2 minor pieces in play and the queen also with nothing moved past its own 3rd rank.
To me it looks like vorex is setting a trap.
The knight's pawn is "poison" in many chess openings like the Poison Pawn variation of the Sicilian.
Hi, G.T., so what would be White's coming next "dominating" move then, showing its initiative?
I don't see any advantange either way for either player. It's still too early to say who is doing what.
smirf,
I just meant there is no "initiative" that any side has won yet. It's only move 6 after all.
I think white doesn't have to defend the pawn, that's all. The move Qxb2 does not have to be defended against. I might be wrong we'll just hafta wait and see.
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6
7. Nc3
I think it's healthy to speculate about upcoming moves but don't criticize the play or use emotional sentiments. It's only a game, and it's being played by 2 programs, not humans.
If you think it's easy, make your own program :)
Maybe poisoned or not - SMIRF takes the Pawn and will lose some tempi in exchange, but hoping the best.
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6
7. Nc3 Qxb2
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2
8. Cd3
It's not always apparent with chasing a pawn is worth the effort. Vortex never gives us what it can't see getting back, sometimes with positional bonus added on for good measure. When it's score is > 0 and it is "down a pawn", for the most part, I know the pawn sacrifice is sound.
But is the score is something like -50 (a pawn is worth 100) then I have some problems determining the value of it. Likewise, when it wins a pawn, but the score is only +50, I know the opposing side is getting something in return for the pawn.
We'll just have to wait and see since the "good stuff" is just beyond the principal variation.
If you take away the silly 5. Bf4 6. Be3 by Vortex, it's playing almost like a human player. I have mixed feelings about just letting the b2 pawn drop. I see no reason why 7. Qc1 couldn't be played. But I also think the b2 pawn is the pawn of least value on the whole board. It don't do squat and it's not part of anything that really matters so letting it fall for 2 free developing moves like Nc3 and Cd3 might make sense, especially since smirf is attacking when the king is not set comfortably and the bishops are doing absolute 0.
Ed, what algorithm let's Vortex drop pawns to get positional counterplay? Can you explain it?
I would have given the hanging B PAWN line a try in an offhand game. I like it!
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6
7. Nc3 Qxb2 8. Cd3 Qa3
SMIRF finally is believing now, to survive that trade with a full Pawn advantage.
I side with smirf player. I don't see how white player gets pawn back. White player with many pieces moved but pawn this is not worth.
Fitz Man : Get real. A Pawn does not worth nothing while we're in Gothic Chess middlegame.
How bad can be the opening play of Gothic Chess computers? This game proves it can be more than i imagined! :-)
White played terrible, yet he is winning.
In its PV-line SMIRF had waited for the black squared Bishop to come back for to protect the b-Pawn. But that would have meant, that this piece would have been moved triple in vain then. As far as I know Vortex, it would never decide for such a move in the opening, even when it would produce the probably best possible position.
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3 Qa3
9. Ne4
Vortex considered 7. Bc1 for about 7 plies, then it switched to 7. d5 Qxb2 8. Bd4 through about ply 11, then on 7. Nc3 Qxb2 8. Qd2 intending 9. Cb1 to threaten black's Queen and get three tempi for the pawn (knight move, chancellor move, queen move) and finally it settled on deploying the Chancellor.
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3
Qa3 9. Ne4 Qa5
I agree with Vortex, that the evaluation has shifted additional 1/4 Pawn units to Black's benefit. The center will see a lot of trades, soon. SMIRF still regards itself to be on the side of the surviver.
So crocodile, what is so bad with the play so far other than the bishop move that white made?
4.Bf3? a stupid move, blocking the f-pawn and the precious archbishop. Archbishop's development is seriously delayed.
6.Be3? no comment needed. Moving a piece twice in the first 6 moves it's a tragedy. And what a surprise! Moves it in the worst square blocking e-pawn too. Just great.
6...Qb6?! Hunting pawns huh? 6...d5 is much better.
7.Nc3! at last a good move.
7...Qxb2? Delaying your opening development in Gothic Chess is suicidal. White wins!
I agree that Bg2 would have been better than Bf3, but blocking the F pawn is not usually an issue in Gothic Chess. The d-pawn in Gothic Chess is "as important" as the king's pawn in regular chess, and the f-pawn, curiously isn't. I've played many games without moving the F pawn through at least 15 moves with no harm.
Cartaphilus, that is because you didn't play with strong enough opponents to exploit your bad play.
f and d pawns are essential factors for controlling the center.
Well I was ranked 1 spot below Bobby Fischer:
http://www.gothicchesslive.com/players-games.php
I think I know what I'm doing.
Someone picking a fight with Cartaphilus, oh, this will be good :)
Carta is one kick ass player dude! I should know, I was known as "BloodOfBulls" on the live game server before that jerk Claude disabled the site when he had a mental breakdown.
And I mowed down my share of players.
http://www.gothicchesslive.com/one-players-games.php?id=430
See for yourself.
This says umbra. For a strange reason i feel it was not Bobby :-)
I also see you are ahead of M Tal. Is this maybe the known Tal? :-)
If you are so good as you say, can you find the solution to white to play and win:
POSITION
Cheers:-)
1. Be4 obviously suggests itself at first glance because white has no other options.
OK then. To make it more difficult i can say that there are 2 moves that are winning.
Can you find both?
By the way, there is a new Gothic Chess server where you can play live. I added the link to the left hand side of the GothicChess.com main page. Also, there is a live chat going on right now.
http://www.dracis.com/chat/ to chat and http://www.dracis.com/wiki/GothicChess/ to play, so if you see this, drop on by!
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3 Qa3 9. Ne4 Qa5
10. d5
Sorry for no diagram I was busy helping with the new live playing site.
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3
Qa3 9. Ne4 Qa5 10. d5 Qxd5
SMIRF is risking a trade of its Queen for White's Chancellor. Currently it is regarding its position to be evaluated as about +3/2 Pawn units.
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3 Qa3 9. Ne4 Qa5 10. d5 Qxd5
11. Cxd5
Vortex thought 10...cxd5 was better. It still sees the score fairly close, +/- 0.08 pawns as it searches. When it is behind by 2 pawns and there is such a score, this usually means something good is going to happen.
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3
Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5
Hmm, SMIRF's evaluation still is growing. There seems to be a fundamental difference between both engines in evaluating this situation.
Here Smirf is wrong. Position is about equal for both sides, with the difference that white is about to gain the initiative.
If he succeeds to do that and black defend incorrectly then game will soon be over.
gothicChessInventor, http://www.dracis.com/wiki/GothicChess/
This site wants to install something on my computer in order to play but my browser says that this something is untrustworthy and not verified.
Have you installed and played there so you can verify its safety, in order to know who to blame if a trojan intrudes in my computer:-)
Does anyone know, whether Greg Strong has written a successor for the current published version 0.9.2 of ChessV or if he would send a renewed version of it to the championship? The actual release seems to be very buggy, whereas it might help here, to have a third working engine's opinion on this position's evaluation.
Hi Crocodile,
Yes I installed the items without any problems. As the author says on the page, "...it makes me sounds like a terrorist..." but his code is completely safe.
Reinhard,
The guy who entered ChessV into the tournmament in the past has not been Greg Strong. This was revealed some time last year. This person still "talks" in such a way to let people think he is Greg Strong for some reason. He never comes out and says things, he lets you "fill in the blanks" and come to the wrong conclusion."
You can always post a note at the discussion board and ask him.
By the way, there are 5 programs in the tournament this year:
Vortex, SMIRF, TSCP Gothic 64, ChessV, and Tornado, a new program from England.
So the ChessV guy, who says in that forum that will kick Vortex's butt is not the programmer of ChessV?
And last year ChessV participated and someone deceived you pretending to be Greg Strong, while the real Greg didn't know anything about it?
LOL!
I think maybe somebody should play the Crocodile a game of Gothic Chess at the new site and see if his play is as good as his mouth is.
:)
Seems like some of us are on at about the same time today. Anybody try out that new Gothic Chess chat yet?
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3 Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5
12. Bxh6+
I think it's safe enough to offer this conditional move to help the game along:
If 12...ixh6 then 13. Nd6
I am online at the chat right now as "GothicChessInventor" in case you couldn't figure out that was me :)
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3
Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6
(if 14. Bxd5 then Cxd5)
Well, so do I with SMIRF's moves. What is Vortex evaluation now about? SMIRF still believes to be +2 Pawn units in front.
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3 Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5 Cxd5
15. Qxd5
Looks like Vortex wants all of these trades.
If 15...cxd5 16. Af3
The score is +0.20 pawns for Vortex after 15. Qxd5. With all of its extension code (to 33 plies) it sees how to get the pawns back.
I'll try and move the PV display area so that the score shows next time. When the PV is more than one line long is pushes the score up in the scroll box.
Both programs have been playing weird. White can force a repetition draw at will if it wanted to.
11...Axd5? is not good and 12. Bxh6+ can't be recommended either. I liked 12. Nd6! right away and of course the pawn capture 11...cxd5 makes the most sense.
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3
Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5
Cxd5 15. Qxd5 cxd5 16. Af3 e6
In the live chat Ed was talking about his recent interview in a French magazine
http://www.onirik.net/spip.php?article3690
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3 Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5 Cxd5 15. Qxd5 cxd5 16. Af3 e6
17. O-O
Vortex at +0.22 pawns now.
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3
Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5
Cxd5 15. Qxd5 cxd5 16. Af3 e6 17. O-O b6
SMIRF insists in an evaluation around +2.2 ...
This would be interesting:
Archbishop against 2 bishops plus 2 pawns!
I predict that obviously the archbishop will prevail when rooks and knight and specifically the white pawns begin to roll.
White will win!
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3 Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5 Cxd5 15. Qxd5 cxd5 16. Af3 e6 17. O-O b6
18. Ah4
Two bishops + 2 pawns are clearly stronger than an Archbishop. That's not what Vortex attributes to its advatange.
Black has 1 piece developed, 2 bad bishops, and neither of its Rooks are a factor right now. Neither the kingside nor the queenside will be safe for black's King, while white might never be placed in check for the rest of the game. The "missing pawns" from white's game are unnoticed, and they might even help white's Rooks get to better posts if the program can make use of them.
There is a strong chance that the sum of all of these goodies should not overpower the material deficeit, but if I were a human playing in this game, I'd prefer the comfort of white's position but wish I had black's material.
Since you can't "have everything", let's see how the game finishes.
:)
By the way, notice Vortex has not moved its e- and f-pawns.
I like the idea of using the blog to have a diagram of a correspondence game in progress. Is there a way to do this automatically? If so, anybody want to play a game with me?
Still it is unclear, which hardware is used by Vortex. I see it always reaching one or two more plys in deep. Since the beginning of this game most people have regarded Vortex as the one and only winner candidate. But SMIRF still goes marching on, watching its evaluation increase. If it would find the time to rearrange its pieces, what could it prevent from winning then?
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3
Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5
Cxd5 15. Qxd5 cxd5 16. Af3 e6 17. O-O b6 18. Ah4 j5
I've noticed in playing Vortex that it disregards pawns from time to time and hunts for strong positions. I don't agree with it's dropping both pawns in the game but I'dve played 7. Nc3 when Qxb2 was an option for black.
Vortex also sometimes does better once it's behind. Sounds weird I know. It's almost like it prefers a material loss of a pawn or 2 if the line that comes next is sharp, win or lose for it.
I think this is because when there's no tactics within its searchable window, it'd rather let lose of pawns to create tactics. It's like it would want to enter a strategically bad position if the stronger side had to defend with care, or "get into the game" when behind is development like black is here.
It's like Vortex plays like a "wreckless Paul Morphy" would if he played Gothic Chess.
I'd like Ed to comment on this and see if it makes any sense.
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3 Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5 Cxd5 15. Qxd5 cxd5 16. Af3 e6 17. O-O b6 18. Ah4 j5
19. Ni5
If 19...hxi5 then 20. Axi5+
I think people have been commenting on what they feel they could do in the position as white, I don't think they meant Vortex could win the various positions with 100% certainty. Most humans say things like "Oh, I can win from here" and that is just human nature.
I also don't think that just because SMIRF has an increasing evaluation then that means it will win. I can't count the number of times Reinhard had mentioned having scores of +3 pawns or more and Vortex still won.
Vortex is Vortex, and SMIRF is SMIRF, they're just 2 computer programs, and they are bound to disagree on evaluations.
I think most people can agree that Vortex makes surprising moves from time to time, moves that are just perplexing, including this move to donate a Knight when it is "already down" in material.
But you have to remember: Vortex has specialized code to generate "only captures" in positons that have them, and it does so to great depths.
For example, look at this position:
http://www.gothicchess.com/images/blog/01_vortex_smirf_blog.jpg
You see the search depth being reported as:
"14-40"
What does that mean? Every move has been generated through 14 plies of nominal depth, and the deepest "all captures" line in the principal variation has been examined to a depth of 40 plies. So, no matter what "quiet" moves exist in the next 40 plies, unless they lead to mates issued by knights, Vortex should not lose any material that is not indicated by its current score.
By that I mean: even the worse line through 40 plies worth of "acceptable captures" cannot possibly create a score worse that what is being reported.
That means, if Vortex "goes down" material, it can recover it. What happens along the way to recovery is what makes the play interesting.
As we have seen a few times in this game, Vortex made some moves that lost material "for no reason." Well, the reason is, it already searched to a ridiculous depth and saw it could get it back, all the while putting the opponent in peril that we have not yet seen.
This is not the "best way to play" the game, but it is rather fun to watch, and I hope to find improvements to this radical approach to the game.
As for the hardware: Vortex runs in 32-bit mode on a 2.4 GHz AMD box. Nothing out of the ordinary, and it is now slow by today's standards.
One of the reasons why I allow you to take 24 hours to make a move is so that you won't complain about "slow hardware". Even if one of us had a machine 10x as fast as the other, the gains made in a 24 hour search would be minimal. With a branching factor of about 3, a machine 9x as fast would only search 3^n = 9; n = 2 plies more than the other one.
With a machine 3 times as fast, the faster would outsearch the slower only about 1 ply more in 24 hours.
So, unless you are running on a machine that is 800 MHz or slower, the 1-ply difference is not because of any hardware differences. It is most likely a software implementation responsible for this.
Cartaphilus, If so, anybody want to play a game with me?
I'm willing to play if you tell me what to do? You will start the blog-topic or me? And how?
Someone should tell us:-)
I think 1 move at least, per day is the best right?
19. Ni5 ????
WOW! Vortex surprised me. Obviously it will respond with h4->h5 and open the position of the King and as I see, it can win the black j-rook perhaps giving his a-rook and then despite being behind in material, since Smirf's pieces would be on the left side corner, Vortex would win the right side pawns of black and march a i-and-j pawn push and win.
Clever plan by...
But you have to remember: Vortex has specialized code to generate "only captures" in positons that have them, and it does so to great depths.
Do you mean quiescent search? What specialized? Every program has it.
Every move has been generated through 14 plies of nominal depth, and the deepest "all captures" line in the principal variation has been examined to a depth of 40 plies.
So, no matter what "quiet" moves exist in the next 40 plies, unless they lead to mates issued by knights, Vortex should not lose any material that is not indicated by its current score.
This implication is wrong of course :-)
Vortex did not examine all, at least 40 plies, capturing sequences. Just a few of them and just the deepest was 40 ply.
So it can lose material in the next let's say 10 (not even 40 you said) plies and to not even have considered about it!
By that I mean: even the worse line through 40 plies worth of "acceptable captures" cannot possibly create a score worse that what is being reported.
This is just wrong as I've said and it would be insane if it happened:-)
But you made it tricky with your ""acceptable captures", so only a god knows what this means :)
With a branching factor of about 3, a machine 9x as fast would only search 3^n = 9; n = 2 plies more than the other one.
How come? What kind of theory is this? Can you explain it to us?
Chessprogramms differ. Evaluation functions are important for that differences. Moreover the decision, whether a position is quiet or not here is very program specific.
When to decide, which of evaluation function candidates would be better to implement, the question is, where to put on that decision. I have not the chance to have thousends of games played for to see, which of those would be more successful. One of my criteria is to prefer a function, which leads to a more continuos and linear development of evaluation values during a game.
Nevertheless there is no guarantee for having a better engine by doing that. I know that there are still a lot of weaknesses in SMIRF, that is why I am planning a complete rewrite as OCTOPUS, which seems more appropriate than to correct or manipulate my first wildly grown chessprogram. But now there is playing just that multiply patched approach, hoping to be sufficiently fit to become number two in the November tournament.
Now SMIRF still is thinking on a game, where Black primarily should be happy to reach a draw.
1. Nh3 g6 {(13.00) -0.164} 2. d4 Nh6 {(12.11) -0.184} 3. g3 c6 {(12.08) -0.191}
4. Bf3 Na6 {(13.24) -0.125} 5. Bf4 f6 {(14.01) +0.334} 6. Be3 Qb6 {(12.00)
+0.531} 7. Nc3 Qxb2 {(15.00) +1.053} 8. Cd3 Qa3 {(13.00) +1.012} 9. Ne4 Qa5
{(14.17) +1.256} 10. d5 Qxd5 {(14.22) +1.941} 11. Cxd5 Axd5 {(15.00) +1.961}
12. Bxh6+ ixh6 {(17.00) +2.266} 13. Nd6 Cxd6 {(17.00) +2.236} 14. Bxd5 Cxd5
{(15.10) +2.266} 15. Qxd5 cxd5 {(14.06) +2.211} 16. Af3 e6 {(16.02) +2.264} 17.
O-O b6 {(15.00) +2.268} 18. Ah4 j5 {(15.01) +2.580} 19. Ni5 hxi5 {(18.00)
+3.107} 20. Axi5+ Ke7 {(18.00) +3.107}
Vortex has more than just a quiescence search. It does what I call extended extensions.
And, I think I know how my own program works :) I did not say it looked at all nodes through 40 plies, I said it looked at all captures from the principal variation forward. It generates moves that are captures, and in the resulting positions, it "pretends" the only legal moves are captures or moves that get out of check or moves that generate checks that can lead to captures. It does an alpha-beta on them, and once there are no such captures or checks, it evaluates them as leaf nodes. This way, you get a "skeletal" tree put into your hash table, and it is a "fast forward" view of your candidate endgames.
A "branching factor" measures at what rate your game tree grows. Look at the nodes reported at Depth "A". When Depth "A+1" completes, divide the total nodes by the node count at "A". This is your "brancing factor." In Gothic Chess, this is usually about 3, meaning for each ply you search, you need to look at 3 times as many positions as the previous ply.
This means two things:
1. It takes you 3 times as long to get to the next ply.
...or equivalently...
2. A machine that is 3 times faster will search 1 ply deeper in the same amount of time as the machine that is 3 times as slow.
So, how long does it take to search 2 plies deeper? The answer is branching factor times branching factor = 3 x 3 = 9.
Meaning it takes 9 times as long to get 2 plies deeper, or a machine 9 times as fast can outsearch a machine 9 times as slow by only 2 plies in the same amount of time, etc.
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3 Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5 Cxd5 15. Qxd5 cxd5 16. Af3 e6 17. O-O b6 18. Ah4 j5 19. Ni5 hxi5 20. Axi5+ Ke7
21. Axh7
From the new diagram, you can see Vortex searched 19-63 plies and returns a score of -82, which is -0.82 pawns (less then one pawn down.) This is an amazing score for 1 Archbishop vs. 3 minor pieces (the difference in the two material distributions) which is closer to being down a piece (-3.0) but Vortex sees it will get almost all of it back with optimal play for both sides.
Also, notice the "DB 7: 2" item. This means positions with 7 pieces are being found, and in two cases it could extend and hit its 5-piece tablebases in RAM. These are "perfect scores" from presolved endgames, so with each subsequent move, this will become a factor.
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3
Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5
Cxd5 15. Qxd5 cxd5 16. Af3 e6 17. O-O b6 18. Ah4 j5 19. Ni5 hxi5 20. Axi5+ Ke7
21. Axh7 Rj6
SMIRF still believes in about +3.1 Pawn units.
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3 Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5 Cxd5 15. Qxd5 cxd5 16. Af3 e6 17. O-O b6 18. Ah4 j5 19. Ni5 hxi5 20. Axi5+ Ke7 21. Axh7 Rj6
22. h4
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3
Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5
Cxd5 15. Qxd5 cxd5 16. Af3 e6 17. O-O b6 18. Ah4 j5 19. Ni5 hxi5 20. Axi5+ Ke7
21. Axh7 Rj6 22. h4 Nc5
With that move SMIRF has begun to reorganize and develop its pieces, what has been delayed because of the early dangerous capturings.
Now this amateur 10x8 and 8x8 multivariant combined engine still holds against a specialized professional program for more than 20 moves and yet regards itself as the leading party.
Nevertheless the outcome still is undefined.
I think Vortex is now playing too risky. The moves are not sound. When you're behind in material, you don't get rid of more material. It dropped pawns for temporary advantages that weren't increased upon. It developed its pieces quickly then suddenly stopped. The attacking didn't begin. It's like getting your army all set to invade, then giving them vacation leave.
When you have someone behind in development you're suppose to open up lines of attack and bring the attack to them before they can complete their development. Something seems to wrong with this new Vortex you have.
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3 Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5 Cxd5 15. Qxd5 cxd5 16. Af3 e6 17. O-O b6 18. Ah4 j5 19. Ni5 hxi5 20. Axi5+ Ke7 21. Axh7 Rj6 22. h4 Nc5
23. h5
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3
Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5
Cxd5 15. Qxd5 cxd5 16. Af3 e6 17. O-O b6 18. Ah4 j5 19. Ni5 hxi5 20. Axi5+ Ke7
21. Axh7 Rj6 22. h4 Nc5 23. h5 g5
Most people tend to think of the endgame as just a "middlegame with fewer pieces."
This is widely incorrect.
In many of my games on BrainKing, people play the Opening like a Grandmaster who invented it, play the middlegame with tactical wizardy, yet 1 move into the endgame, they commit strategic errors that even a 1700 tournament player would not make.
I always find this amusing.
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3 Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5 Cxd5 15. Qxd5 cxd5 16. Af3 e6 17. O-O b6 18. Ah4 j5 19. Ni5 hxi5 20. Axi5+ Ke7 21. Axh7 Rj6 22. h4 Nc5 23. h5 g5
24. h6
Vortex predicted the last move, no sense in delaying.
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3
Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5
Cxd5 15. Qxd5 cxd5 16. Af3 e6 17. O-O b6 18. Ah4 j5 19. Ni5 hxi5 20. Axi5+ Ke7
21. Axh7 Rj6 22. h4 Nc5 23. h5 g5 24. h6 Ba6
The next step to Black's freedom ...
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3 Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5 Cxd5 15. Qxd5 cxd5 16. Af3 e6 17. O-O b6 18. Ah4 j5 19. Ni5 hxi5 20. Axi5+ Ke7 21. Axh7 Rj6 22. h4 Nc5 23. h5 g5 24. h6 Ba6
25. Ag6+
Well you have 8 pawns worth of material in the corner dedicated to stopping 1 pawn on h6.
"Let the freedom ring"
:)
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3
Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5
Cxd5 15. Qxd5 cxd5 16. Af3 e6 17. O-O b6 18. Ah4 j5 19. Ni5 hxi5 20. Axi5+ Ke7
21. Axh7 Rj6 22. h4 Nc5 23. h5 g5 24. h6 Ba6 25. Ag6+ Kd6
If it is working, what is wrong with it?
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3 Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5 Cxd5 15. Qxd5 cxd5 16. Af3 e6 17. O-O b6 18. Ah4 j5 19. Ni5 hxi5 20. Axi5+ Ke7 21. Axh7 Rj6 22. h4 Nc5 23. h5 g5 24. h6 Ba6 25. Ag6+ Ld6
26. e3
Usually when you tie up resources to a defensive take they are being under utilized.
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3
Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5
Cxd5 15. Qxd5 cxd5 16. Af3 e6 17. O-O b6 18. Ah4 j5 19. Ni5 hxi5 20. Axi5+ Ke7
21. Axh7 Rj6 22. h4 Nc5 23. h5 g5 24. h6 Ba6 25. Ag6+ Kd6 26. e3 Be2
Indeed, that is the bad payload of SMIRF having been too late in development for to make risky captures.
Current list of moves with added SMIRF's evaluations:
1. Nh3 g6 {(13.00) -0.164} 2. d4 Nh6 {(12.11) -0.184} 3. g3 c6 {(12.08) -0.191}
4. Bf3 Na6 {(13.24) -0.125} 5. Bf4 f6 {(14.01) +0.334} 6. Be3 Qb6 {(12.00)
+0.531} 7. Nc3 Qxb2 {(15.00) +1.053} 8. Cd3 Qa3 {(13.00) +1.012} 9. Ne4 Qa5
{(14.17) +1.256} 10. d5 Qxd5 {(14.22) +1.941} 11. Cxd5 Axd5 {(15.00) +1.961}
12. Bxh6+ ixh6 {(17.00) +2.266} 13. Nd6 Cxd6 {(17.00) +2.236} 14. Bxd5 Cxd5
{(15.10) +2.266} 15. Qxd5 cxd5 {(14.06) +2.211} 16. Af3 e6 {(16.02) +2.264} 17.
O-O b6 {(15.00) +2.268} 18. Ah4 j5 {(15.01) +2.580} 19. Ni5 hxi5 {(18.00)
+3.107} 20. Axi5+ Ke7 {(18.00) +3.107} 21. Axh7 Rj6 {(19.00) +3.148} 22. h4 Nc5
{(17.01) +3.383} 23. h5 g5 {(18.00) +3.418} 24. h6 Ba6 {(14.04) +3.725} 25.
Ag6+ Kd6 {(14.01) +3.961} 26. e3 Be2 {(12.00) +3.961}
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3 Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5 Cxd5 15. Qxd5 cxd5 16. Af3 e6 17. O-O b6 18. Ah4 j5 19. Ni5 hxi5 20. Axi5+ Ke7 21. Axh7 Rj6 22. h4 Nc5 23. h5 g5 24. h6 Ba6 25. Ag6+ Kd6 26. e3 Be2
27. a4
Vortex likes this position.
This game has been and continues to be interesting. I have to say I wouldn't play like Vortex has in letting go its pawns and its knight and I wouldn't play like smirf has in not getting its pieces in play much sooner.
If these programs both represent the best in the world of Gothic Chess then I think the door is wide open for someone to outplay them in the 2007 Championship.
I've seem older Vortex versions play much more reliably, so I would say to Ed, go back to an older version and stick with it. I think all of the fast forward capturing is messing something up. Sure, it leads to exciting play, but in everybody's chess career we go past the Romantic Era of pure attack and mature as players to resist "going for it". Instead, we play moves that make slow and steady progress and try to accumulate pressure.
I've gone over some smirf games from the site and it does seem like a player that is more steady and consistent. But it seems only to win when the other player loses. I have yet to see it strike out and create something out of nothing, as opposed to Vortex which tries to force something to happen.
I would say here black can win with careful play and proper trades. Now I know I'm not as good as Ed and I'm sure not a programmer like smirf is, but I think I can see what is coming unless a miracle happens.
Other people here seem to have been "Vortex cheerleaders" shouting stuff from the sidelines. The proof is in the play, not in the noise.
smirf seems to be doing better if you ask me, although the reasons are because it has risked less and allowed Vortex to give away too much.
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3
Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5
Cxd5 15. Qxd5 cxd5 16. Af3 e6 17. O-O b6 18. Ah4 j5 19. Ni5 hxi5 20. Axi5+ Ke7
21. Axh7 Rj6 22. h4 Nc5 23. h5 g5 24. h6 Ba6 25. Ag6+ Kd6 26. e3 Be2 27. a4 Bf3
SMIRF has not had the chance to make combinatorical highlights during this game. Instead it seems to be accumulating smaller advantages, where it could get some. My explanation for that is, that though Vortex seems to calculate deeper, SMIRF might have some slightly more precise evaluations.
I thought smirf lost like 20 games against Vortex and only has 1 win. How can you say it has a more precise evaluation routine?
Maybe you have not noticed: SMIRF and Vortex as playing here now are newly released versions, having never before played against each other. Thus my explanations target this one game only. We probably will see more games between both, at least at the November championship. Then we might know better about SMIRF's weaknesses.
Sorry the start of the work week always slows me down
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3
Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5
Cxd5 15. Qxd5 cxd5 16. Af3 e6 17. O-O b6 18. Ah4 j5 19. Ni5 hxi5 20. Axi5+ Ke7
21. Axh7 Rj6 22. h4 Nc5 23. h5 g5 24. h6 Ba6 25. Ag6+ Kd6 26. e3 Be2 27. a4 Bf3
28. Rh2
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3
Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5
Cxd5 15. Qxd5 cxd5 16. Af3 e6 17. O-O b6 18. Ah4 j5 19. Ni5 hxi5 20. Axi5+ Ke7
21. Axh7 Rj6 22. h4 Nc5 23. h5 g5 24. h6 Ba6 25. Ag6+ Kd6 26. e3 Be2 27. a4 Bf3
28. Rh2 Be4
But even no time for a new image of Vortex'es window?
Some idiot from the Czech republic claimed he ordered a Gothic Chess set and never received it, so PayPal shut down our account.
Click on the "shop" link from GothicChess.com and you will see all of the information.
Now because of this internet extorter (he told me to give him a free set or he will send in the fake PayPal claim) I have to get another eCommerce solution up. This has much higher priority.
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3 Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5 Cxd5 15. Qxd5 cxd5 16. Af3 e6 17. O-O b6 18. Ah4 j5 19. Ni5 hxi5 20. Axi5+ Ke7 21. Axh7 Rj6 22. h4 Nc5 23. h5 g5 24. h6 Ba6 25. Ag6+ Kd6 26. e3 Be2 27. a4 Bf3 28. Rh2 Be4
29. Ah5
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3
Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5
Cxd5 15. Qxd5 cxd5 16. Af3 e6 17. O-O b6 18. Ah4 j5 19. Ni5 hxi5 20. Axi5+ Ke7
21. Axh7 Rj6 22. h4 Nc5 23. h5 g5 24. h6 Ba6 25. Ag6+ Kd6 26. e3 Be2 27. a4 Bf3
28. Rh2 Be4 29. Ah5 Bxc2
Sorry for that trouble. I hope, those things will clear up, often caused by impatience paired with communication ability deficits.
I wonder how Vortex is planning to win this game? This is very odd and bad play.
And please don't remove paypal. It's the only way i can pay in any future items i may want.
I agreee, Vortex is unrecognizeable in this game. Letting pawns falls and donating a Knight when down in material. Totally unlike the Vortex from a while ago. Ed needs to change it back.
Another position for solving,
White to play and win:
POSITION
Since no one solved the previous one i must assume that you're totally incapable? :-)
Computers of course like Smirf or Vortex are completely useless here, since they have to search around 30 plies.
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3 Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5 Cxd5 15. Qxd5 cxd5 16. Af3 e6 17. O-O b6 18. Ah4 j5 19. Ni5 hxi5 20. Axi5+ Ke7 21. Axh7 Rj6 22. h4 Nc5 23. h5 g5 24. h6 Ba6 25. Ag6+ Kd6 26. e3 Be2 27. a4 Bf3 28. Rh2 Be4 29. Ah5 Bxc2
30. a5
Rybka was also unpleasantly surprised this week, nothing a code revert can't fix.
Anybody can publish a problem then not offer a solution. Why don't you try backing up you say with your analysis?
White's got an easy win with 1. h7 in the position you showed. Black has to deal with the threat of 2. Rg8 so he'd hafta waste a king move right away. White plays Rg8 anyway and wins easy enough.
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3
Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5
Cxd5 15. Qxd5 cxd5 16. Af3 e6 17. O-O b6 18. Ah4 j5 19. Ni5 hxi5 20. Axi5+ Ke7
21. Axh7 Rj6 22. h4 Nc5 23. h5 g5 24. h6 Ba6 25. Ag6+ Kd6 26. e3 Be2 27. a4 Bf3
28. Rh2 Be4 29. Ah5 Bxc2 30. a5 Rb8
30. a5 Rb8 31. axb6
I switched to the 2004 version of Vortex since somehow I can't find the 2005 and 2005 versions of the code.
That is one ugly board you have there! I can see why you changed the graphics!
What the heck happened in this game? It's like watching 2 drunk people play!
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3
Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5
Cxd5 15. Qxd5 cxd5 16. Af3 e6 17. O-O b6 18. Ah4 j5 19. Ni5 hxi5 20. Axi5+ Ke7
21. Axh7 Rj6 22. h4 Nc5 23. h5 g5 24. h6 Ba6 25. Ag6+ Kd6 26. e3 Be2 27. a4 Bf3
28. Rh2 Be4 29. Ah5 Bxc2 30. a5 Rb8 31. axb6 axb6
What has been wrong with SMIRF? If current Vortex has lost interest to continue this game, we should better end it.
Well smufy smirf I think what the silly boi means is both programs did not do weally weally good moves
;)
Do not forget, that SMIRF has the Black side. It mastered well the task of being behind of White's initial advantage. Watching the evluation flow I am very satisfied with SMIRF's acting. A claimed evaluation of -2.4 (Vortex) after 9 hours of thinking reflects only its pure wish-thinking.
1. Nh3 g6 {(13.00) -0.164} 2. d4 Nh6 {(12.11) -0.184} 3. g3 c6 {(12.08) -0.191}
4. Bf3 Na6 {(13.24) -0.125} 5. Bf4 f6 {(14.01) +0.334} 6. Be3 Qb6 {(12.00)
+0.531} 7. Nc3 Qxb2 {(15.00) +1.053} 8. Cd3 Qa3 {(13.00) +1.012} 9. Ne4 Qa5
{(14.17) +1.256} 10. d5 Qxd5 {(14.22) +1.941} 11. Cxd5 Axd5 {(15.00) +1.961}
12. Bxh6+ ixh6 {(17.00) +2.266} 13. Nd6 Cxd6 {(17.00) +2.236} 14. Bxd5 Cxd5
{(15.10) +2.266} 15. Qxd5 cxd5 {(14.06) +2.211} 16. Af3 e6 {(16.02) +2.264} 17.
O-O b6 {(15.00) +2.268} 18. Ah4 j5 {(15.01) +2.580} 19. Ni5 hxi5 {(18.00)
+3.107 20...Ke7} 20. Axi5+ Ke7 {(18.00) +3.107} 21. Axh7 Rj6 {(19.00) +3.148}
22. h4 Nc5 {(17.01) +3.383} 23. h5 g5 {(18.00) +3.418} 24. h6 Ba6 {(14.04)
+3.725} 25. Ag6+ Kd6 {(14.01) +3.961} 26. e3 Be2 {(12.00) +3.961} 27. a4 Bf3
{(16.01) +4.379} 28. Rh2 Be4 {(16.01) +4.910} 29. Ah5 Bxc2 {(16.18) +4.846} 30.
a5 Rb8 {(15.01) +5.014} 31. axb6 axb6 {(13.05) +5.014}
"Mastered well"?? You'er joking right?
Were you playing over the same game the rest of us were smirf? I didn't see one combination that smirf "won". Vortex donated 2 pawns, then it donated a Knight. These were the obvious mistakes.
I didn't see smirf do anything special, it just played less worse that Vortex did.
You talk as if smirf outplayed Vortex. Nobody here believes that.
Both programs played bad. Vortex obviously has some kind of bug in that version.
Victor, watch the game. SMIRF decided, that the early Pawn capture would work. Vortex did not believe in that. But that early decision defined the following game, where SMIRF gathered small advantages.
Who is claiming, SMIRF had played bad, should specify a situation in this game, where a better continuation for Black would have been on hands.
It was not long ago that I defeated smirf easily enough as seen here http://www.gothicchesslive.com/javascript/game.php?gameid=2229 so trust me when I say, black had much, much better play available. I stopped counting my wins against smirf and the number of times smirf lost on GothicChessLive.com and one game played bad by Vortex won't convince me that it has a great program.
If I know Ed Trice, he is using this game as a rouse to let smirf believe it is better than Vortex, and in the games that count at the World Championship, Vortex will somehow "magically be cured" and destroy the competition. Why do you think Ed set up this game so close to the main event?
Cartaphilus: well, maybe you are right, we will see. SMIRF is not a master player yet, so a master would see passed chances. But the question is, whether Vortex would detect them. Seeing Vortex'es "optimistic" evaluations, I doubt, that this version would be able to find out either. It is very possible, that Ed would send a modified (and improved) Vortex into the coming tournament. As I told already - Smirf is going to become second best, we will see, where it is placing at the end.
Carta,
I wish you would not make statements like that. The 2007 tournament will be played live in front of members of the Main Line Chess Club in Philadelphia. National Master Dan Heisman will be there to observe.
I don't know why you say I create a deliberately weaker version of the program before the match.
Anyway, it's just one game, kind of a "warmup", and now I am using the 2004 version of the program.
31. axb6 axb6
32. f4
Ed,
Are you trying to tell me you didn't let smirf win the game on the wall here http://www.gothicchess.com/wall_of_honor.html just to get him to play in the tournament that year? During a chat on GothicChessLive you told someone "If I don't give smirf some victory he won't play this year because Reinhard is so childish" isn't that true?
To Cartaphilus: there is a lot of noise around of online games - you could see that even here.
Let me assure you, that I have agreed to participate this 2007 event before having started this game here, which still is not finished.
I have no problems with SMIRF to lose a game. As I have repeatedly explained, there had been other reasons to leave a tournament. Well, this one seems to promise a fair performance, thus I did not hesitate to enter a fresh version of SMIRF.
Where is the online site?
Since BlueHorseShoe on here is Claude Jeruchim looking for attention, I suggest everyone ignore him. I was in an argument with Mig Greengard recently, trust me, Kasparov is not posting here.
32...Bh7
33. fxg5
Carta,
I never made that statement on the former GCL site. I never let anyone "win" a game when I operated Vortex. If you check the date of that game, you might even find it occurred after the event you are talking about.
33...fxg5
Since that move was predicted:
34. Af7+
34...Kc6
35. Rc1
35...g4
36. Ag5
36...Rj7
The diagram is wrong silly boyz
;)
I'd have to say smirf looks like it can win from here. It's a strange position though. Vortex has that passed pawn in the H file that smirf has surrounded. Smirf has the B pawn which is passed but Vortex has 2 rooks to keep it from advancing. Pawns on d7 d5 and e6 don't help smirf right now without the doubled pawn on d5 coming off first but then the black king gets exposed. I don't see any combinations waiting to happen so it looks like the material balance isn't going to change any time soon.
Does anybody see some short term goal for black that'll lead to a win? I don't know how to win this game and I'd like to hear from the group.
37. Af4
There you go swimmer girl, a new diagram, are you happy now?
:)
Black can win by trading off all rooks. The black bishop pair can then keep the white archbishop busy while the knight moves around and picks up pawns. Whoever promotes the first pawn will win.
Ra8? now is tricky and i think it draws or wins for white.
After Ra8 there is the deep plan of capturing the g pawn so he will eventually get 2 connected passers g-h and be very dangerous and i don't think black can defend this. Draw in best case for black.
So black has to avoid Ra8? and play Rg8.
37...Rg8 SMIRF's decision ...
Good for it! :-)
You could replay that game e.g. with SMIRF using this PGN:
[Event "Longtime 10x8 Game"]
[Site "GothicChess.BlogSpot.com"]
[Date "2007.09.17"]
[Time "11:10:16"]
[Round "Blog-Game-1"]
[White "Gothic Vortex"]
[Black "Smirf MS-168i"]
[Result "*"]
[Annotator "RS"]
[SetUp "1"]
[FEN "rnbqckabnr/pppppppppp/10/10/10/10/PPPPPPPPPP/RNBQCKABNR w KQkq - 0 1"]
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3
Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5
Cxd5 15. Qxd5 cxd5 16. Af3 e6 17. O-O b6 18. Ah4 j5 19. Ni5 hxi5 20. Axi5+ Ke7
21. Axh7 Rj6 22. h4 Nc5 23. h5 g5 24. h6 Ba6 25. Ag6+ Kd6 26. e3 Be2 27. a4 Bf3
28. Rh2 Be4 29. Ah5 Bxc2 30. a5 Rb8 31. axb6 axb6 32. f4 Bh7 33. fxg5 fxg5 34.
Af7+ Kc6 35. Rc1 g4 36. Ag5 Rj7 37. Af4 Rg8 *
38. Ah5
38...Rj8
So what score does Vortex have now and what is the smirf score?
SMIRF's last evaluation: +6.137
39. Af7
Vortex has the score at -2.76 pawns now.
39...Ra8
40. Rd2
40...j4
41. Ag5
41...Be4
Ed, your statement: "... and now I am using the 2004 version of the program." still is without mentioning a reason for that. Did you prosume your old Vortex to be stronger? Or did you suspect the new Vortex to have a hidden program defect? And is that reason still valid?
He said he couldn't find the source code for the current version of the program that is running, smirf programmer, or can't you read? If he can't look at this source code he can't see if there is a problem, DUH!
To Cartaphilus: he wrote also: "... somehow I can't find the 2005 and 2005 versions of the code." So if I understand that correctly, this is not addressing the code of the current version. Thus there is not any explanation yet for the switch.
I should have said "2005 and 2006 versions" of the source code. I changed them so long ago, and had several machines since then, I can't find what I call the "tested source code." It's around somewhere, probably on a DVD I created.
So, the last "tested" version of the source code I could find was from 2004. I looked at this code, saw nothing wrong with it, so I'll use it.
I forget how the extended extension code works in the 2006 version, so I would need to look at it more closely.
42. Rg1
The extension code is producing an infinite loop, that is why I am switching. Somehow even after 2 hours of searching Vortex could not complete ply 9, something it can do in only a few seconds.
The extended extensions were "going on forever", and not coming back to the root for more nominal depth searching. This is a bug, obviously.
Well, I understand. Good luck in erasing that bug!
Smirf moves: 42...j3
I had several test games at home against the downloadable ChessV release. But it always is crashing, when its King gets in serious attacks. Thus it seems, there are more people which have to fix something in their GC aware programs. SMIRF also is not yet the final version ready for the coming tournament.
43. i3
If I could find the source code I could find the bug! I might have to go back to the 2004 version and recode from memory, not a good thing to try at this point.
The pairings for the Computer World Championship are posted on GothicChess.com now.
I saw the annoncements - and I wondered about the event will be lasting then that long.
In contrast to your Vortex I still have my source code, but I did not want to use it any longer, until I agreed to enter your tournament. Because it is high time to rewrite the whole wildly grewn thing. But now I am putting some vitamines into it instead of letting it stay asleep ...
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3
Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5
Cxd5 15. Qxd5 cxd5 16. Af3 e6 17. O-O b6 18. Ah4 j5 19. Ni5 hxi5 20. Axi5+ Ke7
21. Axh7 Rj6 22. h4 Nc5 23. h5 g5 24. h6 Ba6 25. Ag6+ Kd6 26. e3 Be2 27. a4 Bf3
28. Rh2 Be4 29. Ah5 Bxc2 30. a5 Rb8 31. axb6 axb6 32. f4 Bh7 33. fxg5 fxg5 34.
Af7+ Kc6 35. Rc1 g4 36. Ag5 Rj7 37. Af4 Rg8 38. Ah5 Rj8 39. Af7 Ra8 40. Rd2 j4
41. Ag5 Be4 42. Rg1 j3 43. i3 Be5
So this is the end i guess.
If Smirf of course will not blunder but in this time control this seems improbable.
I would like to see the return match.
Perhaps Smirf may agree to play with me, but that would seem unfair since in correspondence i would crush it :-)
Though Vortex'es analyses might show a less frightening evaluation actually SMIRF's numbers are slowly reaching an 8 Pawn units advantage.
I never have claimed, that SMIRF would play better 10x8 Chess than humans do. I am merely hoping for it to become second in the coming GC championship. SMIRF is my first (multiply repaired and improved) playing chess program and thus needs a total rebuild. Do not forget: SMIRF is a one man amateur project still searching for serious sponsors, helping to make it a professional and competitive program. ;-)
44. Ai4
It make sense to use the blog for correspondence games, but we do have a live chat, and a live playing site. I suggest you play using the chat at http://www.dracis.com/chat/ so that you can save your moves. The online playing module does not save games, so every game you play at http://www.dracis.com/wiki/GothicChess/ will not be saved.
Let me know how it goes.
44...b5
44...b5 is good and wins but 44...Raf8 would make it 0-1 on the spot!
Black pushes the pawn in the B file but why doesn't white get the Archbishop out of the way and push the pawn in the I file? The threat of 2 connected passed pawns must be something Vortex can see, can't it?
I started reading the Rybka discussion again. I notice people have many different piece values. Does this even make sense???
Trice Kaufman Schnarnagl
Pawn 1.0000 1.0000 1.0000
Knight 2.5000 3.0000 3.0000
Bishop 3.0000 3.5000 3.7500
Rook 4.7500 5.5000 5.7500
Archbishop 6.5000 8.0000 7.0000
Chancellor 8.2500 9.5000 9.0000
Queen 8.7500 10.5000 9.5000
Is Trice = Vortex values and Scharnagl = smirf values?
SMIRF uses following average piece values
(without to be overlayed positional impacts):
Q = 9.6005
C = 8.9090
A = 6.8916
R = 5.7112
B = 3.6637
N = 3.0556
P = 1.0000
Those "Scharnagl" values in contrast are memorizable values for human players.
45. Ah5
Vortex uses those number just for the Static Exchange Evaluation portion of its search. It modifies its piece values at different stages of the game.
I forgot how mostly :)
45...b4
46. Axg4
46...d6
47. Rd1
At this point, I might just enter the 2004 version of the program into the computer world championship. It would take too much work to try and do anything with a reasonable amount of assurance that the program would not be corrected.
I don't understand what went wrong in this game for white. 15. Qxd5? and 19. Ni5?? were what we call "unforced errors" in tennis. There wasn't a need to trade Q for C on move 15 and move 19 should get three ???s if there was such a thing.
So why did Vortex self destruct from such a strong position?
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3
Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5
Cxd5 15. Qxd5 cxd5 16. Af3 e6 17. O-O b6 18. Ah4 j5 19. Ni5 hxi5 20. Axi5+ Ke7
21. Axh7 Rj6 22. h4 Nc5 23. h5 g5 24. h6 Ba6 25. Ag6+ Kd6 26. e3 Be2 27. a4 Bf3
28. Rh2 Be4 29. Ah5 Bxc2 30. a5 Rb8 31. axb6 axb6 32. f4 Bh7 33. fxg5 fxg5 34.
Af7+ Kc6 35. Rc1 g4 36. Ag5 Rj7 37. Af4 Rg8 38. Ah5 Rj8 39. Af7 Ra8 40. Rd2 j4
41. Ag5 Be4 42. Rg1 j3 43. i3 Be5 44. Ai4 b5 45. Ah5 b4 46. Axg4 d6 47. Rdd1 b3
Black's b-Pawn promotion tour has been well prepared, while White still was dreaming of vitalizing its neutralized passed h-Pawn.
GothicChessInventor:
47.Rd1 is not a correct move since it is ambiguous. Which Rook to move there?
Correct would be 47.Rdd1 or 47.Rgd1 depending on which move you want of course.
Smirf:
How do you know Vortex intended to play the d-rook to d1? I know it's a better move but still this is irrelevant.
To Crocodile: when in doubt, take the nearest! ;-)
I just saw the 2007 Gothic Chess Computer World Championship schedule. Why is only 1 set of games being played each week? How come the tournament drags on for months?
I never heard anything like this before!
Neither do i. I want to find the reasons for this too.
I mean to start in November and end in April 15th? Just a single game per week?
Stupid decision until we find the reasons.
In 2004 we had humans make moves physically over the board for the programs in the tournament. You can read more about this here:
http://www.chessville.com/GothicChess/ComputerWorldChampionships.htm
The SMIRF program consistently used up 50% of its time before move 10, then it had to move very fast the rest of the game. In one round, SMIRF actually had a forced mate on the board but it ran out of time. Reinhard's posts to his native German discussion board were downright beligerent, basically accusing us of cheating. He knew his time control code was to blame, yet it was easier to claim a human was the reason SMIRF lost rather than the fact the clock expired on move 56 (and, in fact, it went unnoticed until move 57.)
For this reason, I accelerated the Online Website Project, which was not supposed to begin for another 6 months. That way, in 2005, programmers can play their own machines over the net and there would be no accusations.
Of course, SMIRF had disconnect problems, and, of course, this was our fault, not his, since he was the only one disconnecting (again, we were accused of severing his connection.) While other people did sign back on and play their moves after they would (on occasion) disconnect, not SMIRF. He "demanded" an "automatic" re-creation of the position and the clock times, even though this was a very hard thing to code. Still, we took a stab at it, and it was not to his liking either.
Our webmaster eventually grew tired of the ever-enlongating list and he defaulted on his contract rather than support people who did not care for the parts that were working.
But one thing has show true: Reinhard complains about everything.
This year, in 2007, we are only running one pair of games per week so that National Master Dan Heisman, himself a tournament director, can personally watch every move of every game.
That way:
1. Reinhard can't say that there were any humans moving too slow.
2. Reinhard can't say his computer was not as fast as some other computer (all programs will be installed on the same exact machine.)
3. Reinhard can't say there is a disconnect problem.
4. Reinhard will have access to all of the moves played by every program every week.
5. Reinhard will have time to change SMIRF from week to week if he wants to do so.
I think this covers every possible thing that Reinhard could want. And if it doesn't, then "Tough shitsky" as they say in Poland!
:)
47. Rdd1 was correct, and now...
48. Ah5
Well, Ed, I am not interested in putting new life into an old discussion, whether I would be a bad guy or not. So I will concentrate in finishing this game.
48...b2
49. i4
Oh god don't remind me about the whining that went on in the 2005 tournament by Reinhard. What a damn baby!
"I could not reconnect, whah whah whah!"
Good lord man, we were all watching these games, and your behavior was like a 5 year old that pissed in his diaper at Disney World!
Look, people like me want to come here and watch the game. All this name calling should stop. Since that idiot named Claude turned off the online play site because he is an asshole, we should just enjoy the play here without resorting to calling each other names.
Except Claude, because he is a piece of shit.
1. Nh3 g6 2. d4 Nh6 3. g3 c6 4. Bf3 Na6 5. Bf4 f6 6. Be3 Qb6 7. Nc3 Qxb2 8. Cd3
Qa3 9. Ne4 Qa5 10. d5 Qxd5 11. Cxd5 Axd5 12. Bxh6+ ixh6 13. Nd6 Cxd6 14. Bxd5
Cxd5 15. Qxd5 cxd5 16. Af3 e6 17. O-O b6 18. Ah4 j5 19. Ni5 hxi5 20. Axi5+ Ke7
21. Axh7 Rj6 22. h4 Nc5 23. h5 g5 24. h6 Ba6 25. Ag6+ Kd6 26. e3 Be2 27. a4 Bf3
28. Rh2 Be4 29. Ah5 Bxc2 30. a5 Rb8 31. axb6 axb6 32. f4 Bh7 33. fxg5 fxg5 34.
Af7+ Kc6 35. Rc1 g4 36. Ag5 Rj7 37. Af4 Rg8 38. Ah5 Rj8 39. Af7 Ra8 40. Rd2 j4
41. Ag5 Be4 42. Rg1 j3 43. i3 Be5 44. Ai4 b5 45. Ah5 b4 46. Axg4 d6 47. Rdd1 b3
48. Ah5 b2 49. i4 Rji8
ROFL!!!!!!!!!!
Victor OMG that was so damn funny!!!!!!
All of this name calling should stop, except for Claude, because he really is an asshole!
LOL!
Picture of the current game state:
10x8 Image
50. Kh2
50...Rah8
Just a quick note:
If you update your diagram on the 10x8.net site and do not rename it, the blog will automatically show the latest position.
And on another note...
We will have to change the format for the Computer Chess Championship again this year.
A Zillions 2.0 entrant was received, and another program called "Variant Shredder" signed up. I have no idea if this is based on the original shredder program or the author just took creative license when naming it.
51. Af7
Post a Comment