SWJediknight wrote on 01/23/19 at 23:31:57:
Yes, I have seen this theoretical conclusion just about everywhere. But not quite everywhere. The great New York experts in the Two Knights always demurred, e.g. in the pages of Chess Review in the 1950s. In print we usually would read this opinion second-hand via an editor -- they preferred to reveal their analyses OTB only. When they retired from the scene they took their knowledge with them. Maybe the scoresheets and notes which would form the basis of Secrets of the Fried Liver are mouldering forgotten in some New York City basement.

Oh well, that's ancient history. Speaking from personal experience, it's sad to have a lifetime of analysis invested in an opening and have the engine render it irrelevant in the time it takes to reach 30 ply.

Bumping this thread because it's interesting to go back and review the theory with much more powerful hardware 5 years later (almost to the day).

What strikes me is how 9...c5 is just taken for granted as the move to play. Now take Stockfish 10 on a 16-thread CPU, and it finds 9...Be7 is a much better move:

9. f3 Be7 10. cxd4 exd4 11. Qb3 (x-ray on f7) Qf5 12. Ne4 Be6 13. Qd3 0-0 14. Nbd2 Qd5 15. 0-0 Rfe8 16. Rd1 f5 17. Nf2 a6 18. f4 c5 and you have to believe black has excellent compensation here. The bishop pair and lion's share of the board with those connected queenside pawns.

FirebrandX wrote on 12/22/13 at 15:02:30:


Finished my game with the IM:

[Event "Centaur Unrated Theory Game"]
[Site "Chess.com"]
[Date "2013.12.23"]
[White "pfren"]
[Black "FirebrandX"]
[Result "1/2-1/2"]
[ECO "C57"]
[WhiteElo "2330"]
[BlackElo "2240"]
[TimeControl "1 in 5 days"]
[Termination "Game drawn by agreement"]

1.e4 e5 2.Nf3 Nc6 3.Bc4 Nf6 4.Ng5 d5 5.exd5 Nxd5 6.d4 {Thematic Game - This is the starting position.}
Nxd4 7.c3 b5 8.Bxd5 Qxd5 9.f3 c5 10.cxd4 cxd4
11.Nc3 Qc4 12.Ne2 d3 13.Ng3 Qd4 14.Bd2 Be7 15.N5e4 O-O 16.Nf2 Qxb2 17.O-O a5 18.a3 Be6 19.Qe1 Rfc8 20.Rb1 Qxa3
21.Rxb5 f6 22.f4 Bc4 23.Rb7 Ba6 24.Rb1 exf4 25.Qe6+ Kh8 26.Nge4 Bf8 27.Ra1 Qb2 28.Rxa5 Re8 29.Qf7 Qb7 30.Qa2 Qc6
31.Rc1 Qb6 32.Nc5 Bxc5 33.Raxc5 h6 34.Qa4 Re2 35.Qxf4 Qb2 36.Rd1 Qb6 37.Rcc1 Bb7 38.Rf1 Rd8 39.h3 Rd7 40.Rfd1 Qb5
41.Qg3 Kh7 42.Rb1 Qc6 43.Rb2 Qd5 44.Rb6 1/2-1/2

While there were still many moves left before reaching a dead drawn position, we both had other tournaments come up that were taking up our free time elsewhere.  Nevertheless, the final position came after all the fireworks had died out and black was going to hold the draw without much difficulty centaur-wise. Some quick thoughts on the game:

1. It took a LOT of work avoiding losing lines in all throughout the middlegame. During the opening and middlegame phases, pfren got to play natural and obvious moves, meanwhile I had to walk a tightrope and use all my analysis skills to avoid nasty pitfalls the computer was going to walk blindly into.

2. This line just isn't practical at all for black in OTB play. As mentioned in point #1, there's far too much DEEPLY accurate play required of black. Far beyond anything a GM could memorize even.

3. From what I had read researching the opening, Houdini's suggestion of 11...Qc4 hadn't been tried before and black had been losing badly with other lines where the queen hangs back. Based on that, the move looked worthwhile if I was expecting to improve on established theory.

4. Houdini then wanted to follow up with 12...Be7 and flounder away while white builds up winning position. After spending several hours looking at alternatives, I eventually found 12...d3! which I believe was absolutely critical for black's survival. Instead of d3 being a threat to happen, it actually happens straight away!

5. Later in the middlegame, white refused what looked to be an obvious free take of the d3-pawn, but pfren and I both had analyzed that out to be a far easier draw for black. Pfren played 18.a3! instead, which Houdini suggested was the ONLY try for a win. Everything else hastened an easier draw for black. The idea behind a3 is to entice the black queen off a protective diagonal of either the e5 pawn or the soon-to-be e6 Bishop and lose precious time munching the pawn. Indeed pfren felt taking the a3 pawn at all was quite risky, figuring instead that I would plant the queen on a2 behind it to stay on one of the aforementioned protective diagonals to my center. In the end after many hours of branch analysis, I opted for taking the pawn, mostly based on the idea that my only chance to survive was to clear off white's pawns as fast as possible, especially the queenside ones.

6. Once I got in 34...Re2!, I finally felt safe and would be able to hold the draw without having to do intense analysis. The only thing left I had to use human intuition on was ignoring Houdini wanting to advance the kingside pawns towards the end. I knew the draw would be much easier to hold with my pawns still keeping my king safely tucked away.

7. Overall what's interesting to note about this game is the weakness on f2 lasted the entire game. It's actually quite incredible how that one little weakness was the only thing that kept black alive.

sloughter wrote on 11/06/13 at 05:47:31:


I seem to recall analysis presented here in 2007 (?) suggesting that after 1.e4 e5 2.Nf3 Nc6 3.Bc4 Nf6 4.Ng5 d5 5.exd5 Nxd5?!  6. d4! Nxd4?! 7.c3 b5 the move 8.Bd3! is White's most practical route to an advantage.

In the Two Knights Defense (1.e4 e5 2.Nf3 Nc6 3.Bc4 Nf6) 4.Ng5 d5 5.exd5 Nxd5?! yields two main lines, 6.Nxf7, the Fried Liver, and 6.d4, the Lolli.

Jon Edwards in a long article in Chess Life found improvements for White yielding a plus for White in the Fried Liver, but I find the analysis unconvincing (to be covered in another post) and Dan Heisman came up with key original analysis of the Lolli in Chess LIfe suggesting that Black could equalize.

What is clearly bad for Black is 6.d4 Bb4+? 7.c3 Be7 8.Nxf7 Kxf7 9.Qf3+ Ke6 10.O-O & Pinkus has more or less analyzed this as a forced win for White.

Heisman recommends 6.d4 Nxd4! (in my opinion the only hope to equalize) 7.c3 b5 8.Bxd5 Qxd5 9.f3 (While Heisman covers this briefly, it is the critical line in the Heisman variation) c5 10.cxd4 cxd4 11.Nc3! (Much stronger than O-O; if 11...Bb4? 12.a3 +/-) Qd7 12.Ne2 Be7 13.O-O & unless d3 is threat White is just better---lead in development + material edge.

What is especially bad is that Black can never enforce e4 with the pawn on f3 so it is difficult to see how Black creates any real threats with his five to three majority. White threatens Nh3/Nf2 restraining e4. +/=