How is it possible that two World Champions, Correspondence World Champion Dr. Hans Berliner using the strongest computer then in existence who promoted his gambit to a willing chess community and World Champion Garry Kasparov in BCO 2
both miss a check, winning material, and forcing the Black King to move to an undesirable location leading soon to a winning endgame for White?
When I had Fritz 12 look at the main line of the Berliner Gambit, I had hoped to demonstrate the principle of long-term compensation to a student. Unfortunately, when I had Fritz in infinite mode to set up the desired position, it instantly found a cook of the main line, once again, missed by Dr. Berliner and Garry Kasparov. Here is the trivial win for White:
Castle early and often, especially in open position is chess 101, yet in the Two Knights' Defense, castling has been missed in several variations.
4.Ng5 d5 5.exd5 Nxd5 6.O-O?! N (I published this in the Best Question Contest by GM Larry Evans about 30 ? years ago)
5.exd5 Nxd5 6.d4 Be6 7.O-O +/= N? 6.d4 Nxd4 7.O-O +/= N?
5.exd5 Nxd5 6.d4 Bb4ch 7.c3 Be7 8.O-O = (According to GM Lev Alburt) N
For two years I have maintained on line that White was better in the Two Knights' Defense to the scorn of the post members. My claim was that White, in the main line, 5...Na5, that White should not play 8.Be2 which allows Black to open lines or force White to play the awkward Steinitz Variation instead he should play 8.Qf3 instead. The overwhelming belief was that after the move sequence 4.Ng5 d5 5.exd5 Na5 6.Bb5ch c6 7.dxc6 bxc6 8.Qf3 Be7 Black had the better side of equality but---
9.Be2 O-O 10.h4!! This leads to an endgame a pawn up where White can keep his extra pawn indefinitely in a complicated middlegame or Black can stand much worse by keeping the Queens of the board. White anchors the Knight, restrains g5 forever, and castles Queenside with excellent prospects:
A) 10...h6 11.Ne4 & Black has nothing better than 11...Nxe4 12.Qxe4 and White castles Queenside with good prospects of a Kingside attack with g4/g5. +/= to +/
B) 10...Nd5 11.g4! +/-
C) On another post I have demonstrated that on the best try 10...Bg4 11.Qd3 Qxd3 12.Bxd3 Nd5 13.Be2 & Black has no compensation for his pawn deficit and weak Queenside pawn structure.
How do you account for the fact that this simple move sequence was missed by hundreds of post members?
Check your recent thread for my move.
Of course the Homer Simpson Defense loses to 6...d5 7.e5 Ne4-/+
One example of supranormal behavior is that playing c3 to anchor the Knight in the Sicilian Defense is new. When I showed this new idea to GM Lev Alburt he entered an endgame three tempos down that White achieves within 10 moves. Normally that is sufficient for a pawn. I should point out that both Fritz 10 and Lev recommend this move order for Black, but he is playing way too many pawn moves just to get the King to the safe square c2 where the King can never be attacked.
c3 permits White to meet Nxd4 with cxd4 so that White gets two center pawns. It permits the White Queen to get to the squares a4 and b3. It provides a maneuvering square for as many as four pieces, it keeps the Knights out of d4 and b4, it denies Black any counterplay on the c-file by the Rook any activity of the Dragon Bishop. Here is the move order recommended by Lev and Fritz 10.
I set forth the hypothesis that given enough tries I'd defeat Fritz 10 at some time limit from the starting position at move 13 with White on move. The time limit was 60'65 with the position being set up in infinite mode.
How can Fritz 10 make seven blunders in 8 moves?
Nuts! Wrong move order. Here is the right move order.
By "chance" I published the wrong move order!
Lightning can indeed strike twice. Here is a second game from the "Magic" position:
Here is the third game
from the Magic position; this time, though, White is better +/-
Here is the fou
rth game:
Why is Black moving his knight back and forth?
I made the Knight moves to reach the desired position "magic" i.e. with pawns on d6 & e6 for Black, while all of the White pieces occupy their ideal "Classical" layout and White has the center pawns.
The odds are so great that Black shouldn't survive even 10 moves. With correct play by White it is a trivial win. He can play g4/h4, etc. or if White tries d5, if Black replies e5, White wins with Nxe5. This game shows that computers, even strong ones cannot evaluate positions very well. Since their moves are driven by evaluation, this is their Achilles Heel. In both games where Black won, White failed to "strike while the iron was hot" dithered and then sacrificed unsoundly. If computers evaluate improperly, then they will lose against tough opponents.
Human players should avoid any opening where White has a mobile pawn center where White can build on his advantage e.g. the Ruy Lopez, the Queen's Gambit Declined, those variations of the Sicilian where White castles Kingside and gets in f4, or almost all variations of King's Indian, except, perhaps, those sacrifices with c5. The reason is that computers are great at "churning" their pieces to find their ideal squares. Give them pawns to anchor their pieces in an open position and they are deadly.
The Fritz engines have a tough time with gambits sometimes over emphasizing material, or, in the two games here, the initiative. Thus gambits like the Benko, Blumenfeld and other gambits are good choices by the human player against computers.
The kind of position that Fritz screws up is that if I have pawns on h7, g7, f7 and e5 with a Rook on h8 and Knight on g8, should White play Qg3 I will always sac the g-pawn with Nf6 meeting Qxg7 with Rg8.
While it looks as though White only gets 2 tempos for the pawn, in actuality, it is three tempos, because White will have to defend the pawn on g2 for the third tempo in an open position i.e. Black has compensation for the pawn, and, indeed, may get a powerful initiative.
While Nimzovich is highly regarded as a theoretician several of his concepts are subject to challenge. He places great emphasis on pawn chains, yet pawn chains require mutual agreement between players. What is even more telling is that Nimzovich never provided actual examples of pawn chains. If you check his Chapter on pawn chains you will see that every example is a pawn link, not a pawn chain. In fact, the only examples of actual pawn chains are the hypothetical examples proving that Nimzovich knew what a pawn chain was and then promptly forgot what they were.
His whole concept of development is dubious, because he stated that only piece moves count as developing moves. Here is a completely new way to look at development. For beginners this should be second nature, for experienced players it is simply an additional tool. While no substitute for analysis it does make move choices easier based solely on this new system.
Here is the simplest way to define development:
1)Move a pawn one square forward either capturing or moving forward, this counts as 1/2,
2)Move a pawn two squares forward, this counts as 1,
3)Capture and create doubled isolated pawns counts as -1/2
4)Capture and undouble your pawns, this counts as 1,
5)Double a pawn attached to a pawn mass e.g. the exchange variation of the Ruy Lopez this counts -1/4.
6)Capture and open the diagonal of a Bishop, this counts as 1
6)Capture a piece and you gain all the time it takes for the piece to get to the square where it was captured,
7)Any time a piece accesses a square that it cannot access in less than a certain number of moves, then each piece move gains a tempt. Thus the move sequence Nf3/Ng5/Ne4 gains a tempo for each move. However, if the Knight then goes to g3, this counts as a -1 because the Knight can access the g3 square in just two moves i.e. Ne2/Ng3
8)Move a piece to a square that it can access in the same number of moves counts as zero e.g. Bc4/Bd3 counts, respectively as 1, 0.
9)The exception is when the piece retreats to its original square counts as -1.
Here are some concrete examples:
a) Qe2/Qf3 counts as 1,0
b)Qe2/Qe2 counts as 1,1 because the Queen cannot access e3 in under two moves. As a general rule, absent tactics, Qe3 is superior to Qf3 because it gains a tempo,
3)Consider this variation of the Caro-Kann:
1.e4 (1) c6 (1/2) 2.d4 (1) d5 (1) 3.Ne2 (1) dxe4 (1/2) 4.Nxe4 (1) Nf6 (1) 5.Nxf6ch (2---one for developing and one for capturing a piece that has moved once) exf6 (+3 for capturing a piece that had moved three times + 1 for opening the diagonal of the Bishop and -1/4 for doubling the pawns.
Total tempo count White= 6 Black = 6 3/4. Since White has the first move advantage, this is roughly equal.
For the Ruy Lopez, here is the tempo count: 1.e4 (1) e5 (1) 2.Nf3 (1) Nc6 (1) 3.Bb5 (1) a6 (1/2) 4.Bxc6 (2---one for developing and 1 for exchanging a piece that had moved once) dxc6 (2 for capturing a piece that had moved twice, 1 for opening the diagonal of the Bishop and -1/4 for doubling the pawn.
Tempo count: White = 5, Black = 5 1/4. Again White has the first move advantage.
Here is an excellent example of why traditional concepts of development need revision: Consider this variation of the Two Knights' Defense:
1.e4 (1) e5 (1) 2.Nf3 (1) Nc6 (1) 3.Bc4 (1) Nf6 (1) 4.Ng5 (1) d5 (1) 5.exd5 (1/2) Nd4 (1) 6.c3 (1/2) b5 (1) 7.Bf1 (-1) Nxd5 8.Ne4 (1) Ne6 (1) 9.Bxb5ch (1) Bd7 (1) 10.Bxd7ch (2--one for capturing on a new diagonal, one for capturing a piece that has moved once 10.Qxd7 (3)
Tempo count White=8; Black=12 so Black temporarily has a 4 tempo lead in development in a position where the pawn structure is fluid. Naturally, Black stands well.
Note: This counting system does not deal with pawn links and pawn chains.
Going back to the topic, the reason that champions like Fischer, Karpov, Kasparov are able to perform the way they do is because they dedicate their lives to Chess.
They live and breath the game and as a result, anything other than supernatural talent, (bordering on autistic), would be strange indeed.
A change in chess that the Editor of Chess Life said was, "Simple ybut profound?" is an idea to change chess fundamentally without changing the rules or the starting position. What I suggested was that the second player gets the choice of color i.e. he/she can choose the White pieces, and the first player must start the game with the Black pieces. I predicted that with the White pieces that player might play more aggressively. It might make for an interesting psychology experiment.
To the best of my knowledge this has never been suggested before.
The dang thing is on the Fritz!