Supranormal Acitivity in Chess

I would probably play 5...d5 instead of 5...d6, but looking at the final position after 6.d4 this seems fine, black can play 6...c5 and after something like a3 then a5 (either before or after a pawn trade) and far from white having an edge again black is probably already equal (although I tend to favor black a bit but there's probably no real edge for him).

Recently, I got feedback from GM Lev Alburt about a critical ending in the Sicilian Defense:

1.e4 c5 2.Nf3 Nc6 3.d4 cxd4 4.Nxd4 d5 5.Bb5 dxe4 6.Nxc6 Qxd1ch 7.Qxd1 a6 8.Ba4 Bd7 9.Nc3 Bxc6 10.Bxc6ch bxc6 11.Nxe4 f5. Lev said, "I also looked at 12.Ng5 but it doesn't seem to give more than 12.Nd2."

With regard to Nd2 he agreed that best play for both sides is probably, 12.Nd2 e5 13.Re1 Bd6 14.Nc4 O-O-O 15.Bg5 Nf6 16.Ke2 Rhf8 (To avoid doubled pawns) 17.Rad1 Bc7 18.Rxd8ch Kxd8.

"I think that White is slightly better after 19.f3---say 5.5 using my system (What he uses is a ten point system with 10 being a win for White---author) Is it enough to create REAL problems for Black, I don't know."

An important game to illustrate the value of Universal Positions is the game Carlsen-Kramnik, London 2009. Both players started out in Classical mode; then, at a critical juncture, White went into Universal mode. Here is the game score; then I will publish what I think should have happened: 1.c4 Nf6 2.Nc3 e5 3.Nf3 Nc6 4.g3 d5 5.cxd5 Nxd5 6.Bg2 Nb6 7.O-O Be7 8.a3 O-O 9.b4 Be6 10.Rb1 f6 11.d3 a5! 12.b5 Nd4 13.Nd2.

You will note that Carlsen has dropped into a Universal Position, the most difficult to attack. That is why in a gambit opening it is useful to snatch material and then return to a Universal Position with your extra material in tow. Here, Carlsen is using Universal theory to clamp down on the position.

Black should try to return to a Universal Position to defend; the best way to do this is to meet the attack on the b7 pawn with 13...Bc8! White has to be careful here lest he stand worse. if 14.Bb2 as recommended by some players, then Black has the very strong move 14...a4! threatening to plant a piece on the b3 square. Now it is too slow to play 15.Nc4 Nxc4 16.cxc4 Be6 & if 17.Nxa4?? Bb3 -+ so White parachutes the Bishop to b3 with the better game.

You will note that Carlsen recommends 16.a4; I recommend 14.Bb2 a4.

If 14.Nf3? a4! =/+

or 14.e3? Ne6 (You will note that Carlsen recommends protecting c7 with Ne8. Isn't it more active to protect c7 with Ne6 instead?)

15.Qc2 Nc5 (Merely threatening to win the pawn) 16.d4 (what else?) exd4 17.cxd4 & just like that the White Queenside pawns are a shambles =/+

Finally, after looking at the position for about 45 minutes, GM Lev Alburt and I concluded that best for White is 14.Nc4! Nxc4 15.dxc4 Qe8 15.Nd5 Ne6 16.Nxe7ch Qxe7 with about an equal game (Black has a Universal Position)

; in a must win situation Black could try to complicate with: 15...Be6?! (see diagram)

What constitutes "supranormal" behaviour is debatable, but at what point do we draw a distinction between "exceptional" behavior and supranormal behavior? When I proposed that 1.e4 e5 2.Nf3 Nc6 3.Bc4 Nf6 4.Ng5 may lead to a win by White in all variations, I was told it would require an ELO rating of 3600 (I am a Class B postal player). This is what has been accepted by post members either tacitly (not contesting the analysis) or agreeing that it is correct:

4.Ng5 Nxe4 5.Bxf7ch +/=

4.Ng5 Bc5 5.Bxf7ch Ke7 6.Bd5 Rf8 7.Bxc6 dxc6 (bxc6 8.d3 +/-) 8.Nf3 +/-

4.Ng5 d5 5.exd5 Nxd5 6.d4 +/-

4.Ng5 d5 5.exd5 Nd4 6.c3 b5 7.Bf1 Nxd5 8.cxd4 +/-

4.Ng5 d5 5.exd5 b5 6.Bf1 +/-

About the only line left for Black is 4.Ng5 d5 5.exd5 Na5 6.Bb5ch (Bd7 7.Qe2 +/=) c6 7.dxc6 bxc6 8.Qf3! I believe this is stronger than 8.Be2 or 8.Bd3 because White's Queen either occupies or influences the center; one indication that it is potentially sound is that Fritz 8 meets 8.Qf3 with Qd5! offering the exchange of Queens which should be met on White's terms with 9.Be2! & not with 9.Qxd5? Nxd5 = to =/+ due to the threat of Nf4.

After losing five games in a row to ArKheiN, both engaging in human/computer play where he was using a World Champion level computer and I was using Fritz 8. The correct response to 8.Qf3 Be7 9.Be2! (Bxc6ch Nxc6 10.Qxc6ch Bd7 with comp.) 10.h4! The purpose of this move is to anchor the Knight so that Nd5 doesn't hang the Knight, discourage h6 due to the mate threats

 and devalue Black's counter play based on the idea of f5/h6/g5. Here are a couple of games between Moody-Conquistador with this line.

In the game Magic (see diagram below) how long will it take computers to construct the starting position given "routine" programming e.g. ask the computer to come up with the maximum lead in development by White and still have a playable position by a Class B player against a Class B computer (The computer is given one hour/move to crunch the position) when the human player takes less than 2 minutes/move i.e. +/- and not +-?

Would you recognize that given the starting moves 1.c4, 2.d4, 3.e4, that a 1700 computer given one hour/move could beat a World Champion? The reason is that at an hour/move computers are great at crunching positions, but lousy at understanding space. Give a huge spatial advantage, the computer will use this successfully to promote active piece play and the appropriate pawn levers.

Would you recognize that the first move by White in the position is absolutely critical? If the computer had played either 2.e5 or 2.d5, I would have resigned immediately but that after 1...e6 2.Kb1 d6 artificially closes the center. This means that it will take White a very long time to get in f4, a critical move to break up Black's position.

GM Lev Alburt claimed that the starting position was +-; a few moves later, despite no major mistakes by White, he evaluated the position as +/-. He also, on the third move, recommended a blunder by Black because he was following known theory in cramped positions. These theories do not apply to the subject position. Would any chess professionals realized that they had to abandon every known opening principle to play the opening successfully since they would have had to starting thinking on move one.

Would they have realized that White's winning prospects get progressively worse the move development that White gets e.g. that the opening moves 1.e4/2.d4/3.Nf3/4.Nc3 is a lot easier for White to win than the starting position, because White is far more likely to get in either the c4 break or the f4 break opening the position before Black has the ability to defend.

Would you be able to play the entire game using intutition alone and never calculate anything? I came up with a strategy 100 moves deep i.e. I knew that the best hope of winning with Black was to walk my King to a8, not castle and just go after the White King because I knew that the computer would never try to run to the center.

For the analysis of this game see the thread on the Greatest Odds Game Ever?

If constructing the starting position, knowing it was playable against a 1700 computer, knowing that all opening theory was useless and knowing the correct strategy and sticking to that strategy for 100 moves is not supranormal behavior, what is it?

One of the tests of supranormal behavior is for Class players to take the White side of the Magic position and give the Black side to Rybka or Fritz 12.

Here is the way to duplicate my game:

1) Black to move and win,

2)The computer cannot take more than two minutes/move,

3)You may not take less than one hour/move even including obvious moves,

I predict that many players will win with the White pieces.

Conclusion: If this happens as predicted, then I played the Magic position better than Rybka or Fritz 12, computers with ratings of about 2800, when my rating is about 1650.

The only thing is that the computer is not truely playing at 2800 when you give it two minutes per move compared to your hour.  The computer probably cannot get past 12-15 depth when at its strongest should be at 20+.  So in reality, the computer may be playing around 1900 instead of 2800.  In addition, you have the added advantage of having an hour to move.  With all your checking you can raise your level of play by one class, making it possible for you to defeat the computer.

I'm glad someone responded.  It was getting kind of embarrassing -- kind of like I'd stumbled across and was watching someone in flagrante da solo.

"He also knew that if the figures went on rising, no organization, however excellent, could withstand it and that men would come and die in heaps and rot away in the streets, despite the Prefecture, and that the town would see the dead on the public squares clinging to the living with a mixture of justified hatred and ridiculous hope."

Yes, that is this thread.

In the game where I beat a 1700 computer that outrated me, I gave it over an hour/move. I only took a few seconds to a minute to make my moves. In other words, I took less time than I am suggesting that Rybka or Fritz 12 take.

Chess Titans at its top level is around 1700 and I can wipe it off the board pretty easily.  I am skeptical that beating a computer at that level makes you that much better to be honest.

I'd always estimated chess titans about about 1700 too, it's cool to hear someone else say that.

If you give Rybka 5-10 seconds to think it may say it's depth is 15 or 16 or whatever but the quality of those moves is low.  Let it think a few minutes and the depth may not go up much, but the moves will be much stronger.  And this depends on the computer too.  If you have a 64 bit quad core or i7 whatever compared to a 1.0 GHz laptop...  you get the idea, both may read depth=16 after 10 seconds, but the laptop will always lose.

I guess what I mean is that an "old" 1700 computer, Par Excellence or Sphinx Legend in an hour will crunch every legal continuation about 3-5 moves ahead and evaluate them pretty well. When I made just one "logical" move where I actually analyzed the position over the span of several minutes, it crushed me, but using the "perogative" of the exhibitor, I retracted the one bad "logical" move and played strictly by intuition from then on making each decision in a matter of seconds.

Simply put: In a matter of seconds, I was able to avoid any major tactical shots by White 3-5 moves deep and routinely "saw" strategic continuations as much as 10+ moves ahead.

In analysis, though, Fritz 8, in a matter of seconds, did see a key tactical shot missed by the Sphinx Legend at a critical juncture. Fritz sacrificed the exchange to achieve a powerful passed pawn on the fifth, and, at the same time, avoided a back rank mate. Clearly, its evaluation algorithms are much stronger than those of the Chess Titan.

Give Fritz 12 or Rybka, 2 minutes to analyze a move, and I guarantee that they will crunch and evaluate at strong Grandmaster level---and class players thinking an hour/move will beat them with the White side of Magic.

Rybka or Fritz 12 will lose, not because they can't analyze or evaluate the positions well, but because for the first ten moves they will be helpless as a baby trying to figure out how to develop.

Here is the first game I played against Fritz 10 (It was on sale!). It is a critical test of one of the main lines of the Wilkes-Barre Traxler.

Wow your genius clearly needs to be recognized! Sloughterchess for world champion!

It's very impressive thinking strategically 10 moves ahead and spotting tactical shots 3-5 moves deep.

Btw, nice 1400 live rating.

After testing out the various variations of the Bxf7+ Traxler, I think that black is fine after the following moves.

1.e4 e5 2.Nf3 Nc6 3.Bc4 Nf6 4.Ng5 Bc5 5.Bxf7+ Ke7 6.Bd5 Qe8N and every game I have played seems to indicate that this is black's strongest option.

I think that the old main line 6...Rf7 is under fire on more than one front at the moment.

Twits all. next thread.

So...Sofia is the brilliant, lazy one?

 Beliavsky tried 6...Qe8 against Anand in 1991 and lucked out in Vishy's time pressure (rare for him). That game saw 7.d3? d6 8.Bxc6 bxc6? (With 8...Qxc6 Black gets an excellent game e.g. 9.O-O Bg4! when Black is fine.)

Vishy should have met 6.Bd5 Qe8 with 7.Bxc6 dxc6 when we get similar lines to my innovation; White is better.

Beliavshy's 8...bxc6 didn't work out well; the position is +/- after 9.Nc3. Black tried a pawn grab which left him with a losing position.