Computer Chess Club Archives




Subject: Re: Most brilliant novelty from cct7 Witchess-Arasan

Author: Vincent Diepeveen

Date: 08:11:45 02/15/05

Go up one level in this thread

On February 14, 2005 at 17:03:52, John Merlino wrote:

>On February 14, 2005 at 15:57:16, Arturo Ochoa wrote:
>>On February 14, 2005 at 11:40:12, Uri Blass wrote:
>>>On February 14, 2005 at 10:56:24, Vincent Diepeveen wrote:
>>>>On February 14, 2005 at 10:33:12, Jon Dart wrote:
>>>>>A few notes from Arasan's games in CCT7:
>>>>>Game 1 against Homer, Arasan had Black in a QID that Schroer called
>>>>>"a super high-class line, very theoretical". Arasan was in book until
>>>>>move 18. It appears Homer misplayed the next few moves. Arasan's score
>>>>>rapidly climbed and it won.
>>>>>Arasan won easily against Alarm after it blundered here with .. Bxa3:
>>>>>[D] 3q1b1k/1p4pp/rn2rp2/BR2p3/p3N3/P2PP1P1/5P1P/1QR3K1 b - - 0 1
>>>>>Black is not in good shape already, but the pawn can't be taken.
>>>>>Arasan lost against Fafis. The opening was some unusual variant of the
>>>>>Four Knights .. Arasan was out of book at move 7. Arasan's score
>>>>>was positive until move 45. I haven't analyzed this yet so I am
>>>>>not sure where it went wrong but it lost rapidly after that.
>>>>>This game against nullmover gave me some anxious moments. 7 .. Ne8
>>>>>is unusual (..c6 is more common) and Arasan was out of book after
>>>>>that. Black got what looked like a pretty scary k-side attack
>>>>>in the KID. But Arasan defended - in fact its score was never
>>>>>negative. Finally Arasan broke through on the q-side -- standard
>>>>>play in the KID - and won. The nullmover author mentioned his program
>>>>>had no passed pawn code and in general has a simple eval.
>>>>>[Event "?"]
>>>>>[Site ""]
>>>>>[Date "2005.02.13"]
>>>>>[Round "?"]
>>>>>[White "Arasan 9.0"]
>>>>>[Black "nullmover"]
>>>>>[Result "1-0"]
>>>>>[ECO "E87"]
>>>>>[WhiteElo "2594"]
>>>>>[BlackElo "2202"]
>>>>>[TimeControl "3000+3"]
>>>>>1. d4 Nf6 2. c4 g6 3. Nc3 Bg7 4. e4 d6 5. f3 O-O 6. Be3 e5 7. d5 Ne8
>>>>>8. Qd2 f5 9. exf5 gxf5 10. Bd3 Na6 11. Nge2 Nb4 12. O-O f4 13. Bf2
>>>>>Nxd3 14. Qxd3 Rf5 15. Ne4 Rh5 16. b4 Rh6 17. Rfe1 Rg6 18. Kh1 Nf6
>>>>>19. N2c3 Nxe4 20. Nxe4 Bf5 21. Rg1 Kh8 22. a4 Qe7 23. c5 dxc5
>>>>>24. bxc5 Rg8 25. d6 Qf7 26. Rad1 Rh6 27. Rge1 cxd6 28. cxd6 b6
>>>>>29. Qd5 Be6 30. Qd2 Bf8 31. Qc3 Qg7 32. g4 Rh3 33. g5 Bg4 34. Rd3 Bf5
>>>>>35. a5 Rh5 36. Rd5 Bxe4 37. Rxe5 Qf7 38. R5xe4+ Bg7 39. Qc6 Rxg5
>>>>>40. Re8 Rg6 41. axb6 axb6 42. Bxb6 Qa2 43. Rxg8+ Kxg8 44. Re8+
>>>>> 1-0 {nullmover resigns}
>>>>>Against Pharaon, Arasan played a reasonable variant of the Slav and
>>>>>was ok for a long time. Finally at this point Pharaon played Bh6:
>>>>>[D] q6k/3r1p2/p4Pp1/1pRn3p/3PQ3/P6P/1P1B4/6K1 w - - 0 1
>>>>>and then posted the Bishop on g7. Neither Arasan nor Crafty would play
>>>>>Bh6 at the tournament time level on the hardware I have, but Crafty
>>>>>does eventually fail high on it, with a score of +1.7, so this may
>>>>>have been the decisive move.
>>>>>I wasn't watching for a while, but the next time I looked Pharaon was up
>>>>>a Knight--not quite sure how that happened, but seems like it found a
>>>>>nice tactic.
>>>>>Pharaon was strong even before its recent version update and now it
>>>>>is really formidable.
>>>>>In the Chompster game, 37 .. a4 by Chompster was a bad mistake,
>>>>>gifting Arasan with an outside passer:
>>>>>[D] 2q1r1k1/5pp1/5bp1/p7/4PQ2/1Br5/P4RPP/5R1K b - - 0 1
>>>>>But the game got into a bishop of opposite colors ending and was
>>>>>drawn. I actually made the draw manually, which brought a protest
>>>>>from sfarrell: he is right that under the rules this should not
>>>>>have been done without the TD's consent. It seems several programs
>>>>>broke this rule in this round.
>>>>>I was disappointed to lose the last game against cEng (witchess). It
>>>>>had a very unusual opening:
>>>>>[Event "?"]
>>>>>[Site ""]
>>>>>[Date "2005.02.13"]
>>>>>[Round "?"]
>>>>>[White "witchess"]
>>>>>[Black "Arasan 9.0"]
>>>>>[Result "1-0"]
>>>>>[ECO "C28"]
>>>>>[WhiteElo "2397"]
>>>>>[BlackElo "2594"]
>>>>>[TimeControl "3000+3"]
>>>>>1. e4 e5 2. Nc3 Nf6 3. Bc4 Nc6 4. Nf3 Nxe4 5. Nxe4 d5 6. Bd3 dxe4
>>>>>7. Bxe4 Ne7 8. c3 f5 9. Bc2 e4 10. Ne5 Qd5 11. f4 exf3 12. Nxf3 Qe6+
>>>>>13. Kf2 Qb6+ 14. d4 Be6 15. Ba4+ c6 16. Re1 Bd5 17. Bb3 O-O-O 18. Bg5
>>>>>Qc7 19. Bxd5 cxd5 20. Qe2 Qb6 21. c4 Rd7 22. cxd5 Kb8 23. Qe5+ Ka8
>>>>>24. d6 Rxd6 25. Bxe7 Bxe7 26. Qxe7 Rc8 27. Kg1 Rg8 28. Rac1 Rdd8
>>>>> 1-0 {ArasanX resigns}
>>>>>I analyzed this overnight with Crafty but didn't find where Arasan
>>>>>went wrong. I didn't like 7.. Ne7 and 7.. Bd6 seems to be better -
>>>>>this has occurred in a few games with this line. After Ne7, Arasan
>>>>>had its Bishop locked in and failed to develop it.
>>>>I watched this game live and found it a very strong game from witchess.
>>>>Especially because it plays without book. Let's be honest there. That's 700
>>>>rating points (a real strong book).
>>>How did you get that estimate?
>>>Do you have one tournament when a program with no book performed 700 elo worse
>>>than the same program with book?
>>Well.... I love that you continue missing the importance of the opening book. It
>>will mean more easy points for your opponents!
>>I have been reading your same "cantaleta" (*) for years and I have seen how
>>Movei has been beated by books well tuned.
>>Hopefully, you understand that in 20 years. Who knows......
>>(*) Cantaleta = a repetitive nonsense made for years without showing any proof
>>on your behalf.
>I don't think that Uri is saying that a good book is unimportant. There's a BIG
>gap between saying that books aren't important and saying that they're worth 700
>points. What Vincent is saying is that if a top engine was playing without its
>well-tuned book it couldn't beat a master  -- and this is ludicrous.

It is not.

750 rating points is about rounded off the 100% chance.

700 points means that if you join a world champs without a book you have 0%
chance to win the title.

Even a year or 6 ago, I remember Marty Hirsch + Necchi estimating having an
average book versus good book already at 300 points difference.

I'm speaking about no book versus very good book.

>Preparing for the CM vs Christiansen match two and a half years ago, I played a
>bunch of games on ICC with The King on a relatively humble P4-2.4, just for fun,

Blitz haha.

Your proof is 0.

Please show up without a book with chessmaster in a world champs and see how you
will get CRUSHED.

In fact chessmaster never shows up there. I bet the lack of a good book...

>without a book. It easily beat every human that it came up against. It even did
>well against most engines including accounts running Ruffian and Crafty, and
>even occasionally beat Chess Tiger.
>The CM account's rating is now in the high 2700s, right around where Chompster
>is. So, since Chompster tied for 8th in score in CCT7, if you give it another
>700 rating points for having a good book, it will be in the high 3400s on ICC
>and would easily have beaten any other engine in the CCT?? -- Uh, I don't think

You use online ratings. That is what i call online toying.

If you play WITHOUT book online at standard, then people will at a certain point
find how to win from your program and repeat that 5000 times until you give up
and start using a book.

So please don't quote complete nonsense here.

You are not even *near* to understanding how important it is to have a very good
book versus no book.

>Now, a good book is obviously better than no book, and I would be hard-pressed
>to put a rating number on it. But 700 points is silly.

about 1000 points online makes more sense as they will repeat the same line 100
thousands of times, let alone in a world championship computerchess.

If you lose second round, you lose the 4th too and the 6th and the 8th and so



This page took 0.03 seconds to execute

Last modified: Thu, 07 Jul 11 08:48:38 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.