13....Be7 14. Bf2 Nxb2 !? 15. Qb3 Nxd4
16. Qxb2 Nxe2+ 17 Nxe2

Would you post the rest of this game, which I don't have?  I looked at 17...Qd1+ 18.Be1 Bc5+ 19.Kh1 Qxe2 20.Qxe2 Bxe2 21.Nxe6 Bc4 22.Nxd8 Rxd8 23.Rxg7 Bd4 24.Rc1 Bxe5 25.Rxh7 Bxa2 26.Bg3 Bxg3 27.hxg3 c6 28.Re1 Rd7 29.Rxd7 Kxd7 30.Ra1 Bd5 31.Rxa7 c5.  I have no idea if Black can hold this.  It looks difficult, but perhaps there is some hope.  Obviously there are alternatives for White along the way.  I'm not sure if Black has any, except maybe 30...Be6 31.Rxa7 Kc7.

"Are we sure that 13...Be7 is no good?2

In the books, it is considerable worse than Qe8. But I think, in order
for me (us?) to be convinced we should look at it as well.

I will post a few (just a few) things, perhaps even today.

As a follow up to my previous message, here is something else

I mentioned before that I was not completely happy with 14...Be7
after

1. e4 Nf6 2. e5 Nd5 3. c4 Nb6 4. d4 d6 5. f4 dxe5 6. fxe5 Nc6 7. Be3 Bf5 8. Nc3 e6 9. Nf3 Qd7, 10.Be2 0-0-0 11.0-0 Bg4 12.Ng5 Nxc4 13.Rxf7 Qe8 14.Nb5!?

(the line with 14 .... Nxe3 has been analysed before). Here are some lines after 14....Be7 15 Bf2

A) 15....Bxe2 16. Qxe2 Rd7 17. Rxg7 Bxg5 18. Rxd7 Qxd7 19 Qxc4
and White has 1 pawn more. This may not be a sure win but i think
White is better (18 Rxg5 Nb6 has a similar assessment)

B) (This was the line that has been bothering me but perhaps
there is an improvement) 15...Nxb2 16 Qc2 Bxe2 17 Qxe2 Rd7
18. Rxg7 Bxg5 19. Rxd7 Qxd7 20. Qxb7 and black does not appear
to be worse - compare with similar line given in A where White has
Queen on c4 but still has a pawn on b2). Similar with 19 Rxg5 Na4
and similar to a position discussed in A - however, Black has the KN on
a4 instead of b6 and White doe not have the b2 Pawn. So, this appears to suggest that Black is doing Ok.
However, it appears that 18 d5 ! is an improvement. Play can follow
with 18... Rxd5 19 Qxb2 a6 20 a4 (Nxc7!? perhaps should also be
analysed) axb5 21 axb5 Nb8 22  Qc1 and White appears to
have an attack. However, I think this should be deeply analysed
before making a final asssesment of JW's move 14 Nb5

Analysis Game:
Four Pawns Attack with 6) .... Nc6 9) Nf3 Qd7  10) Be2 O-O-O   16) Qb3 Bb4:
Verdict: Strong Advantage White in the end game.

     Black is OK in the middle game, but white seemingly is able to seize the initiative in an unbalanced end game.
Specialist minor piece technique was required to win the highly interesting endgame, which involved two knights v rook.
The game involving moves 1 to 22 was previously analysed by Markovich and others.


1) e4 Nf6  2) e5 Nd5  3) c4 Nb6  4) d4 d6  5) f4 dxe5  6) fxe5 Nc6
7) Be3 Bf5  8) Nc3 e6  9) Nf3 Qd7  10) Be2 O-O-O (I had previously
thought that this dashing long castling move should be better than
10) … Rd8, but unfortunately the converse seems more likely.)


11) O-O Bg4 12) Ng5 Nxc4  13) Rxf7 Qe8  14) Nb5 Nxe3  15) Rxc7+ Kb8

16) Qb3 Bb4  17) Bxg4  Nxg4  18) Nxe6 Qh5  19) h3 Ngxe5

20) dxe5 Qxe5  21) Rxc6 bxc6  22) Nbc7 Rd4  23) a3! (Markovich mentions  

23) Na6+ etc.)  23) ….Re4  24) axb4 Re1+  25) Rxe1 Qxe1+  26) Kh2 Qe5+  

(26...Rc8 27.Na6+ Ka8  {27) … Kb7  28) Nac5+ Kb8  29) Qc3 +-} 28.Nec7+ Kb7 29.Nc5+ Kxc7

(29...Kb8 30.Qc3!  Qxc3 31.N7a6+ Ka8 32.bxc3) 30.Qf7+ +-)


27) Qg3 Qxg3 (Forced exchange of queens)

28) Kxg3 Kb7  29) Kf4!±  White has a shattered queen side pawn structure, but Black’s position is gloomy due to the following reasons.

1) The black rook has restricted activity. The rook is not able to occupy a square,  
   which would lead to control of a file.
2) White’s pair of knights is very powerful and they control key squares.
3) The white king is a powerful attacking piece and black’s king side pawns cannot
   be adequately protected.
4) Isolated double pawns are usually very weak, but white’s mangled queen side
   pawns are relatively safe. The black rook is currently contained and the black king  
   cannot approach the white’s weak pawn structure due to the a5, b5, c5 barrier,
   which is constructed by the b4 pawn and the knight at c7.

The move 29) Kf4! was overlooked in favour of 29) b5 with my chess program.
I guess it is still advantageous to have an understanding of endgame theory!

29) …. Rc8   30) Ke5 Kb6
30) ….Rxc7 31.Nxc7 Kxc7  Unfortunately, white will win the pawn grab and pawn promotion race.
32.Ke6 Kb6 33.g4 Kb5 34.g5 Kxb4 35.Kf7 g6 36.Kg7 c5 37.Kxh7 Kb3 38.h4 Kxb2 39.h5 c4 40.hxg6 c3 41.g7 c2
42.g8Q c1Q 43.Qb8+ Ka2 44.Qxa7+ Kb2 45.Qb7+ Ka1 46.g6 +-)

31) Kd6 g6  32) g4 g5  33) Kd7 Rg8  34) Ne8 h6  35) Nd6 Rh8  
36)Ke7 Rh7+   37) Nf7 h5  38) Nexg5 hxg4  39) hxg4 Rh2 40) Ne5 Re2  41) Ngf3 Kc7  42) g5 Rxb2
(At last, the rook can grab the weak queen side pawns, but it is just too late)  
43) g6 Rxb4  44) Ng5 Rc8 45)  Nef7 Rb7+  46) Kd6 Rb6 47.Ne6 c5+ 48.Kd5 Rb8 49.g7 Kb7 50.Nfd8+ +-


However, not every avenue of the variations arising from 10) Be2 O-O-O has been totally analyzed. It would be certainly be fantastic
if such a variation could be played with strong winning prospects.

It is peculiar that it is not black, but white who must demonstrate with great virtuosity
of handling of the knight pair!

An overextended machine analysis from a winning position is perhaps unnecessary, but it was quite instructive in observing how
the knight pair can be used to win with great effect against the rook.

If someone knows about the fundamentals or a literature reference on how to play with two knights, I would be interested to know.

Some general advice was mentioned in the book Winning End Game Strategy by Beliavsky and Mikhalchishin.
The game Keres-Szabo 1953 (ie. before the age of computers) was used as a reference.

The advice was
a) Get them together
b) Get them to co-ordinate.


Referring to an earlier thought in Markovich’s earlier posting: Was Watson’s omission of an evaluation on this line an implication that the line was in trouble?
I was going to say that this is a strong possibility, but after being updated by recent posts, things are not so clear?!

lg wrote on 03/20/09 at 18:58:04:

Is it the thread, or just your posts?  I see five white stars on my posts, you have four yellow ones on yours.  Apparently I'm an admiral, and you're only a ship's captain.

lg wrote on 03/20/09 at 16:01:19:


Nor do I, at least, I try not to.  I fully agree that good moves are often best spotted with the unaided eye.  I merely say, yes, I do use silicon when analyzing this stuff.

Thanks lg, I will look at 20.a3 in III-b.  I understood that your evaluation of the line in question was equivocal, but it seemed to me that Watson was suggesting that Black was in a bad way.
 
In case there is doubt, I meant that my analyses are machine-assisted.