Capítulo 5. DISEÑO DEL ELECTRODO
5.1 Generalidades
5.1.1 Métricas y criterios de seguridad
or what one horse power is capable of
There are many horses trained to help their riders so as
to run upon any one that appears with a drawn sword ...
t the end of the 1 6th century the French was probably taken on trust. I, on my own experience, have been fortunate enough to satisfy myself of the unusual capabilities of a knight.
Gheorghiu - Yusupov Luzern 1 985
The Romanian player went in for this position , erroneously assuming that he would be able to construct an impregnable fortress.
45 ...
46 'it>e2
f4!
If 46 gxf4, then 46 . . .<�:'ld6! 47 f3 (or 47 i.d2 lbf5 48 h5 'it>f6 49 .te1 'it>g7) 4 7 . . . gxf3 48 'it>e3 lLlf5+ 49 '>t>xf3 lbxd4+ 50 'it>g4 lbxb3, and Black wins.
M ichel Montaigne
46 ... lbd6!
As Florian Gheorghiu informed me after the game, he overlooked this knight move in his adjournment analysis. 46 .. . 'it>f5 would not have given anything because of 4 7 'it>d3, while if 46 .. . lbf6 there would have followed 47 f3!.
47 'it>d3
Black's task would have been more compli
cated after 47 f3. It would appear that 47 . . . gxf3+ 48 '>t>xf3 lLlf5 49 'it>xf4 lbxd4 th rows away the win , since White activates his bishop: 50 .tf2 lbxb3 51 .tb6 lbd2 52 .txa5 b3 53 i.c3 lbc4 54 g4 b2 55 i.xb2 lbxb2 56 a5, or 51 . . . lbc1 52 'it>e3! lba2 53 .txa5 b3 54 Wd2 h5 55 .tc3!. I nteresting play results if instead of 54 . . . h5 Black plays 54 . . . d4!? 55 g4 'it>d5. In reply 56 'it>d3!?
lbc1 + 57 'it>d2 comes i nto consideration, but White can also go in for a sharp variation suggested later by Mark Dvoretsky: 56 g5 'it>e4 57 h5 d3 58 i.c3! ttJxc3 59 'it>xc3 'it>e3 60 g6 d2 6 1 g7 d 1 'if 62 g8iV 'it'c2+ 63 'it>b4 b2 64 'ifg3+, and the king will hardly be able to avoid perpetual check.
The correct continuation is 47 .. . lbf5! 48 fxg4 ttJxd4+ 49 'it>d3 lLlf3 50 .tf2 lbe5+ 5 1 'it>e2 (if 51 'it>d2 there follows 5 1 . . . ttJxg4 52 .tb6 fxg3) 51 . . . f3+ 52 'it>f1 (similar variations occur after 52 'it>d2 lbxg4 53 .tb6 'it>e5 54 .txa5 'it>e4 55 .txb4 f2 56 'it>e2 d4) 52 . . . lbxg4 53 .tb6 'it>e5 54 .txa5 'it>e4 55 i.b6 (or 55 .txb4 '>t>e3 56 i.e 1 f2 57 .txf2+
Solo for a Knight
ctJ
1 03I'Llxf2 58 a5 d4 59 a6 d3 60 a? d2 6 1 a8'ii d1'i'+ 62 �g2 'i!Vh 1 mate) 55 .. . d4 56 a5 f2!
57 'it>g2 (57 a6 'it>f3 58 ii.xd4 lDh2 mate) 57 ... d3 58 a6 d2 59 a? f1'iV+ , and White loses.
47 ... lbts
Now White is in zugzwang, and he himself is forced to break up his fortress.
48 hS
48 �d2 would also not have saved White in view of 48 .. .fxg3 49 fxg3ltJxg3 50 �f4ltJf5!
51 !Ji.c7 g3.
48 ...
49 fx g3 50 h6
fx g3
�f6
White's last hope is the vulnerable placing of the black pawns on the queenside. Thus the careless 50 .. . ltJxh6?? is answered by 5 1 ixb4!.
50... <;;t>g6!
Not so convincing is 50 .. . �e6 5 1 �f2 lDxh6 52 �e3 with chances of a d raw. When he made this move, Black had to calculate the variation which occu rred in the game.
51 �d2 ltJx g3
52 .tx b4 ax b4
52 ... ltJe4? 53 �xa5 g3 54 �c7.
53 aS
head start: it requires just three more moves, whereas the black knight can reach the aS-square only in fou r moves. Disap
pointment awaits Black if he tries to queen his own pawn: 53 .. . ltJf5? 54 a6 g3 55 a? g2 56 a8'ilt' g 1 'ii 57 'ilt'g8+. However, as we know, a well-trained horse is capable of u nusual feats . . .
53 ...
54�e3
lDhS!!
54 a6 ltJf4+ 55 �e3 lbe6 56 a? CDc? and wins.
54 ... lbf6
55�4 �x h6
56 a6 ltJd7
57 a7 lbb6
58�x g4 �g6
59�4 �
60�g4 ltJa8
White resigned .
Yusupov - Li Zunian Luzern 1 985
In this position the game was adjourned for the second time. Although during the first adjournment session I managed to win a pawn thanks to the enthusiasm of the white knight, which accomplished an heroic raid in In the race to queen, the white a-pawn has a the enemy rear -
ltJg4-f6-g8xh6-g8-e7-1 04
�
Solo for a Knightc6-d4-e2--g3 - a draw still seemed to me to be the most probable outcome. After a brief analysis it transpired that apart from the exchange of the g-pawn there was no other realistic plan of playing for a win . After this White is left with a single target - the e4-pawn . The impression was that Black could fairly easily solve the problem of its defence.
However, serious work on the position inspired hope, and I began to realise that the last white piece, standing modestly at g3, was truly a 'Montaignian' knight.
59. . . 'it>e5 60 'it>h5 f6
61 g5 fx g5
62 'it>x g5
Black is at the crossroads, since the bishop can defend the pawn from various sides. For a long time the plan chosen by the Chinese player also seemed the strongest to me.
62... �c6
Black keeps his bishop on the b7-a8 squares, and when his king is evicted from e5 it aims for d3.
63 tt:Jts �as
Of course, not 63 .. . 'it>d5 because of 64 tt::le7+ .
64 tt::le7!
White must prevent the passage of the black king to d3. For example, 64 tt::lh6? leads to a
d raw after 64 .. . 'it>d5 65 'it>f4 Wc4 66 "2lf5 'it>d3.
6 4 ... 'it>d6 The more accurate 64 . . . �b7 will be ana
lysed later.
65 tt::lg6 'it>d5 66 'it>f4
Wh ite's plan takes shape. The winning idea is to occupy the key e5-square with the knight. From there it not only covers the d3-and c4-squares, but also aims for d7 or f7.
66 . . . 'it>c5
In the event of the natural 66 .. . Wc4 White would have won by 67 tt::le5+! 'it>c3 68 "2ld7!.
The threat is 69 tt::lc5, after 68 . . . �c4 there is the fork 69 tt::lb6+ , while if 68 . . . 'it>b4, then simply 69 tt::lf6 .
67 tt:Je5! �b7
Or 67 . . . 'it>b4 68 tt::ld7, and Black loses.
68 tt::lf7!
Since there is no satisfactory defence against the threat of 69 tt::lg5 (if 68 . . . 'it>c4 69 tt::ld6+ ), Black resigned .
Let us return to the position after 64 ttJe7.
I nstead of 64 .. . 'it>d6 Black had the more cunning 64 .. . �b7. If now 65 tt::lg6+ 'it>d5 66
�f4, then 66 .. . �c4 67 tt::le5+ 'it>c3.
This is a position of mutual zugzwang. With Black to move he would be lost: 68 .. . �d2 69
Solo for a Knight
lZJ
1 05ti:ld7, 68 . . . i..a8 69 lZ'ld7 '01tc4 70 lZ'lb6+, or 68 ... '01tb4 69 tt:Jf7 '01tc3 70 lZ'ld6. But it is White to move and he is u nable to win : 68 ti:lf7 'it>d3, or 68 lZ'ld7 '01tc4 69 'iti>e5 i.c6.
The natural 66 '01tf4? was a mistake; White can win by 66 '01tf5!! '01tc4 (66 .. . i..c8+ 67 '01tf4 ib 7 68 lZ'le5 i.a8 69 lZ'ld7, and Black has no defence) 67 lZ'le5+ '01tc3 68 '01tf4, and the situation analysed by us is reached , but with Black to move.
It remains to clarify what would have happened if Black had carried out his plan more accurately, i.e. obtained the position in the last but one diagram with his bishop on aB (with the bishop on b7 White wins by 1 ti:lf5, and if 1 . . . '01td5, then 2 '01tf4 with the irresistible threat of 3 lZ'lg3). In this case 1 ti:lf1! leads to a win . After 1 . . . 'it>d5 there follows 2 lZ'ld2 'it>e5 (3 '01tf4 was threatened ) 3lLlc4+ '01te6 (3 . . . '01td5 4 lZ'lb6+) 4 'it>f4 and 5 ti:ld2, while if 1 . . . i..b7 - 2 tZ'lh2 '01td5 3 lZ'lg4 'it>c4 4 lZ'le5+ 'iti>c3 5 '01tf4 , and again a familiar position is reached (cf. the last diagram).
Thus the system of defence with the bishop on b7-a8 runs into a far from obvious
refutation .
Black could have stuck to another line of defence with his bishop on g2-h 1 . We will examine the following important positions.
Here White wins irrespective of the turn to move.
1 lZ'lg6+ '01td5 2 'it>f4 i.f1
If 2 . . . i..e2, then 3 lZ'lf8! i..f3 4 lZ'lh7 '01tc4 (lZ'lf6+ was threatened ) 5 lZ'lg5.
3 tt:Je 7 + '01te6
3 . . . 'it>d6 loses immediately in view of 4 lZ'lg8!
i..g2 5 lZ'lf6 . 4lZ'lc8!!
This at first sight ridiculous move becomes u nderstandable, if the goal of the knight's u nusual route is noticed - the c3-square.
4 ... i.d3
Other continuations also fail to save Black:
a ) 4 . . . i.g2 5 lZ'lb6 '01td6 6 lZ'la4 '01td5 7 lZ'lc3+;
b) 4 . . . '01td7 5 lZ'lb6+ '01tc6 6 lZ'la4 and 7 lZ'lc3;
c) 4 . . . i..b5 5 '01txe4 '01td7 6 lZ'lb6+ 'it>c6 7 lZ'ld5;
d ) 4 . . . 'it>d5 5 lZ'lb6+ '01tc5 6 lZ'le7+ 'it>d6 7 lZ'lf6 . 5lZ'lb6 i..c2 (otherwise lZ'lb6-a4-c3) 6lZ'lc4 Although White has not in fact managed to transfer his knight to c3, his achievements are very considerable: the black bishop has been forced onto the b 1 -h7 diagonal, where it is less well placed .
6 ... 'it>d5 7lZ'ld2 i.d3
8 'it>f5
Black is in zugzwang and is forced to allow the knight to go to f1 (8 .. . i.e2 9 lZ'lxe4 i.d3 1 0 f3).
1 06
�
Solo for a Knight8 ... �c2 9 tL'lf1 �d1
Or 9 . . . �c4 1 0 tL'lg3 �d5 1 1 tL'lh5 and wins.
10 tL'lh2 �c2 11 t2Jg4 'it>c4 12 tL'lf6 �d3 13 t2Jxe4 �e2 14 �f4, and White wins.
Here I should make a slight digression and refer the reader to the start of this interesting endgame, where the author praises the 'enthusiasm of the white knight' . In order to eliminate the last bulwark of Black's defence - the e4-pawn , the white knight had to complete a veritable round-the-world jour
ney (f4--g6--e 7 -c8-b!H;4-d2-f1 -h2--g4-f6-e4).
I n the position from the last but one d iagram it could have been Black to move.
1 ... Sl.f1 2 tL'lg6+ �d5 3 �4 Sl.g2 4 tL'lh4!
Sl.f1
If 4 . . . .th 1 , then 5 �g3! �c4 6 �h2 , forcing the exchange of bishop for knight.
5 tL'lf5 Sl.g2 6 t2Jg3
A very important position of mutual zug
zwang . If it is Black to move he is forced to occupy the f3-square with his bishop, allowing the knight to go to f1 .
6 ... Sl.f3 7 tL'lf1 Sl.d1
7 . . . Sl.g2 is bad because of 8 tL'ld2 ! Sl.h 1 9
�g3.
8 tL'lh2 Sl.c2 9 �f5
If 9 t2Jg4, then 9 . .. �e6.
9 •.• �c4 10 tL'lf1! �d5 ( 1 0 .. . 'it>d3 1 1 tL'lg3) 11
tL'lg3 followed by tL'lh5, transposing into a
winning position which is already familiar to us.
Although the bishop is at g2, all the same Wh ite is able to win .
1 tL'lf5!
Nothing is given by 1 tL'lh5 .tf3! (but not 1 . . . .t h 1 2 t2Jf4 .tf3 because of 3 tL'lg6+ '.t>d5 4 �f4 .th 1 5 tL'lh4 �c4 6 �g3 �d3 7 '.t>h2
�e2 8 �xh 1 �xf2 9 t2Jf5) 2 t2Jf4 .t h 1 ! . 1 ... �d5
Or 1 . . . .tf3 2 tL'lh4 .td 1 3 tL'lg6+ �d5 4 'ii'f5 Sl.f3 5 �f4, similar to the main variation.
1 . . . Sl.h 1 2 tL'lh4! is bad for Black.
2 tL'lh4! .tf1 3 �f5! �e2 4 tL'lg6 Sl.f3 5 �f4 .tg2
Black also loses after 5 . . . .te2 6 tL'le7+ �e6 7 t2Jf5 �d5 8 tL'lg3 .tf3 (the position of mutual zugzwang from the last but one d iagram has been reached ) 9 tL'lf1 etc.
6 tL'lh4 Sl.f1 7 t2Jf5 .tg2 8 tL'lg3
Agai n a familiar mutual zugzwang position.
White wins.
It is more difficult to win when Black moves first:
1 ... Sl.h3! (preventing the important manoeu
vre t2Jg3-f5-h4) 2 �h4! (White tries to give
Solo for a Knight
C2J
1 07his opponent the move) 2 ... .i.c8
Or 2 . . . .i.g2 3 'it>g4 .i.f3+ 4 'it>g5 .i.g2, reaching the position from the last diagram with White to move.
3 'it>h5! .i.d7
If 3 . .. .i.h3, then 4 'it>g5, and White has succeeded in giving his opponent the move.
4 . . . i.c8 5 lt:Jh5 .i.h3 6 lt:Jf4 followed by lt:Jg6+
and r.t>f4 transposes i nto variations analysed earlier.
4 'it>g6! .i.g4
Or 4 . .. �c8 5 lt:Jh5 Wd5 6 tt:Jf6+.
5 lt:Jh5 .i.f3
If 5 . . . �h3, then 6 lt:Jf4 is possible.
6 lt:Jf4 .i.h1 7 'it>g5 ..tf3
This position has already occurred i n the notes. I will remind you of the winning method : 8 lt:Jg6+ 'it>d5 9 'it>f4 .i.h 1 1 0 lt:Jh4 'it>c4 1 1 'it>g3 'it>d3 1 2 'it>h2 'it>e2 1 3 'it>xh 1 'it>xf2 1 4 lt:Jf5.
With this the author concludes his analysis of this interesting endgame. I will be very indebted to readers for any corrections, refinements or refutations.
1 08
�
Mark Dvoretsky