Skype coaching session, 21 May 2017.
Please note: Re-structuring of the sessions now offers the option for guests to
bring their own games for group discussion.
Recordings of some previous coaching sessions can be found at:
http://www.open-aurec.com/Skype/PaulBenson/PaulBenson.htm
The following game was examined, the games 51 moves is given with annotations
and then again without annotations.
Paul Benson.
* * *
Annotated game.
White: N. Dzagnidze, 2429.
Black: R. Danielmeier, 2112.
Event:Gibraltar Chess Festival 2008.
Result: 1-0 in 51 moves.
Opening: Gruenfeld Defence, Exchange Variation, Nadanian Attack, D85.
1. d4 Nf6
2. c4 g6
3. Nc3 d5
4. cxd5 Nxd5
5. Na4 e5
| Distinguishing between Tal and Neanderthal is not so easy as one might think.
This very active move is aiming for simplifications from the black perspective.
Somewhat paradoxical, as anyone playing the Gruenfeld Defence must be prepared
for those lines with, 5. e4 Nxc3, followed with the challenging black, pawn c5,
leading to a sharp middlegame. Perhaps black is worried about the 300+ Elo
rating point difference and has veered off into a system which transforms what
could / should have been the middlegame into an endgame in embryo. There is
nothing wrong with5. ... Bg7, or even, 5. ... Nb6. ||
6. dxe5
| Almost everyone with white seems to take this pawn, and why not?
Instead, 6. a3 e4, turning it into a peculiar French Defence Advanced Variation
with colours reversed has been seen when, 7. e3 Bg7 8. Qc2 O-O 9. Bd2 Nc6 10.
Ne2 Re8 11. O-O a5 12. Nec3 Bf5 13. Nxd5 Qxd5 14. Bc4 Qd6 15. Bc3 Reb8 16. Bb5
Na7 17. Be2 b5 18. Nc5 b4 19. axb4 axb4 20. Be1 Nc6, black was doing fine and
won in 34 moves. ||
6. ... Bb4+
| Initiating a startling forcing sequence designed to create a pawn imbalance
or two, or three.
Instead, 6. ... Nc6, is solid after which, 7. Nf3 Ndb4 8. Bg5 Qxd1+ 9. Kxd1 Be6
10. Nc3 Bg7 11. a3 Nd5 12. e4 Nxc3 13. bxc3 Nxe5 14. Nxe5 Bxe5 15. Kc2 h6 16.
Be3 O-O-O 17. f4 Bd6 18. Bd3 f6, chances seem to be equal, led to a draw in 45
moves. ||
7. Bd2
| White should not seek tactricks with, 7. Nc3 Nxc3 8. Qb3, hitting the two
black pieces, as, 8. ... Nxa2+, leaves white to choose between the joys of, 9.
Qxb4 Nxb4, losing the queen for a bishop, or, 9. Bd2 Qxd2+ mate. Black in the
game now plays a most astonishing concept, combining imagination, tactics,
positional play, and perhaps grabbing some time on the clock the first time
this move appeared on a board. Find this one without external assistance of any
sort and you should treat yourself to a pint! ||
7. ... Ne3
| Yes, that is, Ne3. We are now entering a sequence of moves which once started
seems to offer little option for varying from the commonly played choice. It
might be worth taking some time here working out just what black has in mind
with this apparent beginner's blunder.
If black does not trust this bizarre, Ne3, option, there is the more solid, 7.
... Qe7, when, 8. a3 Bxd2+ 9. Qxd2 Nb6 10. Nxb6 axb6, and black is fine. ||
8. fxe3
| Declining with, 8. Qc1, permits black to simplify with, 8. ... Qxd2+ 9. Qxd2
Bxd2+ 10. Kxd2 Nc4+ 11. Kc3 Nxe5, and black is comfortable. So with the game
choice white has tripled e-pawns for the lost black knight. Hardly sufficient
positional compensation for a piece, which tells us there must be some tactics
lurking still to be found. ||
8. ... Bxd2+
9. Qxd2
| The white queen has been deflected from an important task, what is she no
longer doing by sitting on d2 which she was doing on d1? The surprising, 9.
Kf2, has been tried once before, but Grand Rabbit sniffs a possible Grand
Master mutual pre-game unwillingness to fight, that game was drawn in a few
more moves. ||
9. ... Qh4+
| A little more than a check for the sake of it. Noticing tricks up and down
the files seems fairly easy, noticing subtleties across the full width of the
the ranks is easy to overlook. ||
10. g3
| Not just blocking an annoying check but preparing to fianchetto the f1
bishop. ||
10. ... Qxa4
| Ah, now the tactics makes sense. The black knight dragged the f2 pawn to the
e3 square and thus opened up the e1 - h4 diagonal. A bishop trade with check on
d2 deflects the white queen from her defence of the white a4 knight. Then comes
a black queen check on h4 to pick up the hapless horse. Tactics should be easy
when you know how.
Perhaps a more important question is why have both players been content to
allow this forcing sequence?
Grand Rabbit offers some thoughts as to the motivations which led to this
position along with possible future aspirations.
White has snaffled a pawn along the way, but since it is essentially the e5
pawn, it is unlikely to survive sustained pressure from black.
A pair of minor pieces have been traded and it is only move 10, not boding well
for lovers of middlegame complexity.
Furthermore, as the game proceeds, it seems that white will struggle to keep
queens on the board.
This implies that black is making strong progress in heading for an ending
where it is white who has pawn weaknesses.
Maybe so, but white is going to find some play as the pieces develop,
especially that long h1 - a8 light square diagonal.
So who is better?
Wrong question.
One should instead ask who has the easier task of finding an active plan and so
have the opportunity to take the initiative. ||
11. Bg2
| Instead, 11. Qd4 Qa5+ 12. b4, and black has tried, Qa6, or, Qb6, or, pawn c5,
with mixed results. More testing clearly required before judgement can be
passed here, or perhaps ask Fritz for an independent opinion? ||
11. ... O-O
12. Qc3 Na6
13. Nf3 Qb4
| Black makes his intentions perfectly clear.
The variation beginning with the pawn thrust, 5. ... e5, is designed to remove
as much middlegame complexity as possible at the temporary cost of a pawn.
Now comes an offer to trade queens which white cannot escape.
The game is about to enter a curious limbo.
A pair of rooks and minor pieces each is not complex enough for a middlegame
but not yet simple enough to be an endgame. ||
14. Rc1 c6
| White to play and pluck a move from the Grand Masters mysterious
yet-to-be-published handbook -
"The ABC Of Making Something Look Something Like It Isn't". Opponents
confronted with such mysteries usually fall into three categories before
replying:
(A). Amusement.
(B). Bemusement.
(C). Confusement.
If you can work out just what white has planned for the next few moves, then
award yourself a much-deserved Grand Rabbit Norm. ||
15. h4
| Yes, this move fits well with ideas of an assault against an opposing
fianchettoed king.
Usual plan, open up the h-file and then pile in with all the heavy pieces
you've got.
There is however just a little problem.
That is, queens are about to be traded off, removing virtually all chances of
an all-out kingside mating attack.
Have another think why, 15. h4, is of value in the mind of a Grand
Master-to-be. ||
15. ... Be6
16. a3 Qxc3+
17. Rxc3 Bd5
| Preparing the trade, Bxf3, to be played once the white e5 pawn has been
attacked sufficiently to ensure it can be won. That elusive Grand Rabbit Norm
is still available if the next white move and subsequent strategy can be found.
Clue: Strategy can be both short and long term, and since at the time of the
coaching session it was a General Election for the U.K., perhaps think of who
is going to have majorities and where. ||
18. h5
| Perhaps this move wasn't all that difficult to find? However there will be no
G.R. Norm if you did not find the precise reasons for white choosing to play,
pawn h5. ||
18. ... Rfe8
19. Rh4
| Grand Rabbit wanders warily out from the warren to try and explain some Grand
Master-to-be thinking.
Aha, so the plan of, pawn h4 - h5, was to allow the white h1 rook access to the
centre via an unusual route, was it?
A quick move count reveals the standard procedure of say, O-O, and, Rd1, and
then, Rd4, takes 3 moves to place the rook in the centre of the board.
So why choose, pawn h4, and pawn h5, and, Rh4, which if the rook is to hit the
centre with perhaps, Rd4, costs 4 moves?
Perhaps because from h4 the rook can control the entire white 4th rank, meaning
it is not just the d4 square is available?
Not very convincing, so there must be another and perhaps better reason for
wanting a white pawn on the h5 square.
The conventional plan, O-O, Rd1, Rd4, does nothing about those all important
imbalancing pawn majorities.
White has a centre / kingside 5 - 3 pawn majority, though this is quickly going
to be reduced to 4 - 3 when the e5 pawn falls.
The advance of the h-pawn is designed to further reduce the centre / kingside
pawns with a, hxg6, trade, ultimately intending that white be left with a
single passed pawn on the kingside.
Please be honest with yourself now, did you work out all the strategy for the
advance of the white h-pawn?
If so, take a G.R. Norm. ||
19. ... Bxf3
| Black decides to regain the pawn at the earliest opportunity.
Instead either, 19. ... Rad8, or, 19. ... Kg7, leaving white guessing when the
threat will be executed was also worth considering.
Black seems to be in a hurry to trade units as soon as possible when perhaps
other moves making developmental gains might have been advisable.
Simplifications are fine providing positional losses are not accumulating while
trades take place.
Initiative over material, until there is no more initiative to be gained, then
by all means regain the material as due reward for patience. ||
20. exf3
| Remember those tripled pawns which black was hoping to exploit in the
endgame? It seems if black wants to regain a pawn, most likely the e5 pawn,
then a black, Bxf3, capture was needed, unless black could somehow arrange the
knight on the queenside to trade on f3, not easy to imagine. This means the
regaining of the pawn by black cleans up the white pawn structure into a mobile
4 - 3 majority. ||
20. ... Rxe5
21. hxg6 hxg6
| White now has a cleaned-up mobile centre / kingside 3 - 2 majority ready to
start moving, black has a queenside 3 - 2 majority yet to find a plan. ||
22. Rd4 Rae8
23. Kf2 Rb5
| This rook probe gains nothing substantial. Instead black should mirror the
white king movement toward the centre with, 23. ... Kf8, with thoughts of using
the king to seek rook trades on the d-file when appropriate, after which the
king will at least be centralised for the forthcoming knight against bishop
ending. ||
24. Rd2 Nc7
25. e4 Re7
26. f4 Ne6
| White has taken some central space and is about to give the g2 bishop options
again up the h1 - a8 diagonal. While there are definitely thoughts of nibbling
away at those juicy morsels on c6 and b7 from the g2 square, is there another
possible option for this bishop which might give black a few problems to solve?
||
27. e5 Rc5
28. Rxc5 Nxc5
| The black manoeuvring has removed a pair of rooks, fine, but in the meantime
white has been making positional gains. Black should have instead spent time
centralising the king before seeking exchanges of rooks at the earliest
opportunity. Compare the locations of the kings when the remaining rooks are
traded and decide which king is better and why. ||
29. Ke3 Rd7
| At the risk of annoyance, 29. ... Kf8, might be an idea worthy of
consideration. ||
30. Rxd7 Nxd7
| The black campaign of eliminating material as quickly as possible seems to be
succeeding. Move 30 and there are only 5 pawns and a minor piece each left in
play, should be easy to further reduce material to achieve that much sought
after draw, shouldn't it? ||
31. Kd4 Kf8
| At last, but black really should have centralised the king before seeking the
trade of all rooks.
Though come to think of it, Grand Rabbit is retro-wondering why did black want
to trade rooks at all?
Were those white rooks located on c3 and d2 really monsters in need of
eradication?
Not really, they could not improve their position by force.
It seems black has fallen into the age-old trap of thinking trading material
will make the game easier to handle.
Not automatically so.
Sometimes it is better to keep material on the board as one's own pieces can
then create problems for the opponent if he/she become too ambitious and try to
create activity which cannot be fully justified.
Yes, it is true that a much stronger opponent trying to weave a web of
complexities can be unnerving.
However, it is well known that stronger players will enter into, "Casino-Mode",
when they find themselves confronted with a position where safe and solid
strategy will not make progress.
Depending on one's perspective, their play either evolves or devolves from,
"Calculated Gambol into Calculated Gamble".
This risky style of play is based on experience.
If the statistics reveal sufficient wins over losses then the gamble is
justified in their mind.
Here we shall never know if white was willing to turn, "Gambol into Gamble",
because black has set out on a campaign of reducing material at every
opportunity.
Now we have a bishop against knight endgame where all the positional plusses
favour the bishop.
Then a quick glance at the ratings reveals a 300+ Elo point gap in favour of
the Grand Master-to-be... ||
32. Bh3
| An important move. This bishop is not going to find any fun on the h1 - a8
diagonal and so instead creates pressure on the h3 - c8 diagonal, a much more
fertile furrow. ||
32. ... Ke7
| Important decision for white - Should she play, 33. Bxd7, taking the game
into a king and pawn ending? ||
33. b4
| Grand Master-to-be says, no king and pawn ending, thank you. After 33. Bxd7
Kxd7 34. Kc5 b6+, the white king will not be able to enter the black queenside.
The advanced nature of the white centre / kingside pawns would be of no value.
White can indeed reduce the centre / kingside 3 - 2 majority to a single passed
pawn, but the black king simply occupy the square in front of the white pawn
and then creates a passed pawn of his own on the queenside.
Perhaps some discussion on the forthcoming bishop against knight ending might
assist?
General features, in no particular order of importance, include:
When there are mobile pawns on both sides of the board, a centrally placed
bishop will be able to influence both regions, making it a better piece than a
knight.
A bishop can lose a move by sliding forwards or backwards along lines of
influence.
This means it can keep pointing in an important direction, make a move, but
retain all the threats it had before the move was played.
A knight can never make a move and keep all previous influences intact.
Attacks and defences are transferred from one colour complex to the other.
This flexibility of colour-transference is a disadvantage if a player with a
knight is trying to hold a position by simply doing nothing.
Specific features of the position, again in no order of importance, include:
The white pawns are on dark squares and so can only be attacked by the knight
from light squares, meaning it is the job of the bishop to defend those light
squares which the white king cannot reach.
The white king has chances of entry into the black queenside via the d6 square,
this reduces the black king into the less desirable role of defender.
The centralised white king must play a part in keeping the activity of the
black knight in the centre to a minimum.
There will be occasions when the bishop versus knight stand-off will occur.
Each piece will be mutually restricting the other, it might seem like
neutralisation, but this is rarely so.
Whichever piece in the stand-off needs to move, but cannot move safely, will be
the worse of the two.
The black knight will have a strong location if it can establish itself on a
dark square outpost, but just how can the black pawn formation set one up?
The white bishop has a potential entry on the c8 square, which would force the
black queenside pawns to advance.
If this occurs, then possibilities of the white king entering through released
squares needs to be prevented.
White has a dynamic, pawn e6, potential fork, depending on where the black king
sits, and if a trade of e-pawn for f-pawn occurs, the black g6 pawn will then
be an isolated target for the white bishop.
If the game should ever be reduced to white only having the light square bishop
and the a-pawn, then white will win as the bishop can take control of the a8
promotion square.
Lots to take in there.
It seems the Rule of Wood-Chopping contains a curious inversion:
"Simplifications Create Complexities." ||
33. ... Nb6
| When on d7 the black knight blocked all attempts of the white bishop to land
on the c8 square. Now on b6 it backwardly guards the c8 square. If black can
keep this arrangement up with, Nd7 - Nb6 - Nd7 - Nb6, then how can white make
progress? ||
34. Bg4
| The bishop wastes a move while keeping whatever influences / opportunities it
has along the h3 - c8 diagonal. This alone is not enough to disrupt the black
knight b6 / d7 defensive shuffle. So just what can white do to break the black
repetitive defence? ||
34. ... Nd7
35. a4
| Aha, white intends to meet, 35. ... Nb6, with, 36. a5, closing down the black
knight shuffle. Black is fearful of any movement of the queenside pawns, as
this will either create a weakness for the bishop to probe at, or permit the
white king an entry. These threats yet to happen compel black to lash out, just
what white was hoping might occur. ||
35. ... f5
36. exf6+
| Clearly white is not having the bishop shut out of play.
This trade of pawns can only favour white.
Sure, the black knight can now become an attacking unit, but it only has one
sensible outpost and the white king is centrally placed to limit opportunities.
There is also the little matter of the e5 square now being available to the
white king, which means the black king is going to have to work much harder to
prevent incursion.
Lastly, the g6 pawn is now isolated and as luck would have it, sitting
immobilised on a light square. ||
36. ... Nxf6
37. Bc8 b6
| The black queenside pawn formation is being forced into advancing, and as
they do so, more squares will become available for the white king. White should
also now be thinking in terms of creating the bishop versus knight stand-off
when available, this is going to be an invaluable means of putting a squeeze on
the black position. ||
38. Ke5 Nd5
| For the sake of definition, this placement of the knight on d5 constitutes an
outpost. It is however vulnerable to undermining and white wastes no time in
demonstrating this point for us. ||
39. b5 Nb4
| Aggressive intentions with, 39. ... Nc3, fails horribly for black. White wins
comfortably with, 40. bxc6 Nxa4 41. c7, and the black king cannot prevent
promotion of the c-pawn. The white idea is simply, 42. Bg4, to be followed
with, 43. c8=Q. After 41. c7, black must give up the knight with, 41. ... b5
42. Bg4 Nb6 43. c8=Q Nxc8 44. Bxc8, and the black queenside pawns can be easily
contained by the bishop beginning with, Ba6 - Bc4 - Bb3, after which white then
plays on the kingside as if it is a king and pawn ending.
There is an inspirio-desperado idea for black which might have been worth a
try. The craziness begins with, 39. ... cxb5, offering up the black d5 knight
for a queenside pawn mass. If white declines the black knight by recapturing
with, 40. axb5, then, 40. ... Nc7 41. Ba6 Nxa6 42. bxa6, and black has a
protected passed b-pawn, this is an easy draw.
Instead after, 39. ... cxb5, if white takes up the challenge with, 40. Kxd5
bxa4, white is set the problem of rounding up the black queenside pawn mass
while maintaining a pawn over on the kingside. Looks easy to assess, does it?
One could spend many happy hours exploring a multitude of variations and still
not be certain if the correct judgement has been reached. Happy hunting! ||
40. Bb7
| Perhaps not so easy to recognise with the b-file cluttered with pawns, but
this is a bishop versus knight stand-off. That is, the opposing minor pieces
are on the same file / rank and separated by two squares. Keep this minor piece
stand-off pattern in mind, white will need to use it again to restrain the
black knight before the white king can then make the crucial queenside
invasion. ||
40. ... c5
| This protected passed pawn will not be given the chance to influence the
game. White to play and give up one bishop versus knight stand-off pattern only
to create a different one. ||
41. Be4
| It never rains but it pours. A few moves back black lashed out with, pawn f5,
in order to give the knight a route into the white queenside. However this pawn
advance left the g6 pawn undefended and white now takes full advantage by
re-grouping / retreating the bishop with a very useful gain of tempo. It is not
just sculling and tug-of-war where one makes progress by going backwards. ||
41. ... Kf7
| Necessary to protect the g6 pawn, but now black has lost control of the
important d6 square, the white king will soon invade the black queenside. White
to play and resist the obvious and instead make a very useful defensive move.
Clue: All stand-offs are equal, but some stand-offs are more equal than others.
||
42. Bb1
| The retreat to the b1 square is a neat way of playing a move without
significantly altering the position. This particular bishop versus knight
stand-off is very restricting for black, the knight has not a single safe move
available. ||
42. ... a6
| Black needs to eliminate all the queenside pawn, even if it means losing a
pawn in the process.
This would lead to a draw, the knight and g6 pawn would be sufficient to hold
the white kingside pawns.
Instead leaving the queenside untouched needs careful consideration, 42. ...
Kg7, white can invade with, 43. Kd6.
Ah, but surely the white king advance has let the black c5 pawn have a free run
to force white to give up the bishop?
Yes, play might continue, 43. ... c4 44. Kc7 c3 45. Kb7 c2 46. Bxc2 Nxc2, and
sure enough, white has lost her bishop.
But in return the white king can mop up the black queenside pawns with, 47.
Kxa7, when there is no means for black to protect the b6 pawn.
The black knight will then be helpless against a pair of advancing white
queenside passed pawns, white will quickly win this one. ||
43. bxa6 Nxa6
44. Kd6 Nb4
| The black c-pawn is ready to sprint to win material, but again it will be far
too slow. ||
45. Kc7 Nd5+
| And the proof, instead, 45. ... c4 46. Kxb6 c3 47. a5 c2 48. Bxc2 Nxc2 49. a6
Na3 50. a7 Nc4+ 51. Kc5, and the black knight cannot cover the a8 promotion
square, white will win this. ||
46. Kc6
| White demonstrates that no matter hard they might try, a knight and pawns
cannot enter into an unbreakable mutual protection scheme. ||
46. ... Nc3
| Gaining a tempo of attack on the white b1 bishop does nothing to protect the
doomed black b6 pawn. ||
47. Bc2 g5
| Instead trying for kingside clearance with, 47. ... Ne2 48. Kxb6 Nxg3 49. a5,
fails as the black knight is stranded over on the kingside, the white bishop on
the b1 - h7 diagonal prevents the black knight from re-entering the queenside.
||
48. Kxb6 gxf4
49. gxf4 Nd5+
50. Kb7
| So many ways to win here, white is adopting the self-imposed standard of
finding the most efficient route to victory.
Instead, 50. Kxc5 Nc7 51. Kb6 Nd5+ 52. Kb7, transposes to the game with black
minus the passed c5 pawn.
Why is the king being placed immediately on the b7 square?
So it guards the b6 and c7 squares, making it impossible for the black knight
to give itself up for the advancing white a-pawn. ||
50. ... Nc3
| Had black instead tried, 50. Ne3, white would have needed to be alert to the
attack on the bishop and respond with, 51. a5, letting it go. After, 51. ...
Nxc2 52 a6 Nb4 53. a7 c4 54. a8=Q c3 55. Qc8 c2 56. Qc4+, white has inflicted a
zugzwang on the black queenside.
There are several ways to win here, perhaps simply marching the king into the
kingside, assists the advance of the passed f4 pawn, until the black king is on
the back rank when the queen can become active again to join in for a mate.
One could also retreat the white king down the b-file to dislodge the black
knight from protecting the passed black c2 pawn.
Another option to the game choice is, 50. ... Nb4, when the materialistic, 51.
a5 Kf6 52. a6 Nxa6 53. Kxa6, is an easy win for white.
Finally there is, 50. Nxf4 51. a5 Ne6 52. a6, when black would like to play,
Nxc5+, and remove the final white pawn. Unfortunately capturing your own pawn
is illegal, and strongly hints why the white king refused to capture the black
c-pawn when it was previously available.
And for completeness there is, 50. Nxf4 51. a5 Ne6 52. a6 Nd8+ 53. Kb6, and the
black knight has no means to prevent the white a-pawn from promoting.
Aren't endgames easy when someone else is playing them? ||
51. a5
Black resigns, 1-0
| Black has a few final tries which all fail:
(A). If, 51. ... Nb5 52. a6 Ke6 53. a7 Nxa7 54. Kxa7, with a decisive advantage
for white.
(B). Or if, 51. ... Nb5 52 a6 Nd6+ 53. Kb8 Nb5 54. Bd3, and the black knight
cannot prevent the a-pawn promotion.
(C). Or if, 51. ... Nb5 52 a6 Nd6+ 53. Kb8 Nc4 54. a7 Nb6 55. Kb7, and yet
again the black knight cannot prevent the a-pawn promotion. ||
* * *
## Unannotated game.
White: N. Dzagnidze, 2429.
Black: R. Danielmeier, 2112.
Event:Gibraltar Chess Festival 2008.
Result: 1-0 in 51 moves.
Opening: Gruenfeld Defence, Exchange Variation, Nadanian Attack, D85.
1. d4 Nf6
2. c4 g6
3. Nc3 d5
4. cxd5 Nxd5
5. Na4 e5
6. dxe5 Bb4+
7. Bd2 Ne3
8. fxe3 Bxd2+
9. Qxd2 Qh4+
10. g3 Qxa4
11. Bg2 O-O
12. Qc3 Na6
13. Nf3 Qb4
14. Rc1 c6
15. h4 Be6
16. a3 Qxc3+
17. Rxc3 Bd5
18. h5 Rfe8
19. Rh4 Bxf3
20. exf3 Rxe5
21. hxg6 hxg6
22. Rd4 Rae8
23. Kf2 Rb5
24. Rd2 Nc7
25. e4 Re7
26. f4 Ne6
27. e5 Rc5
28. Rxc5 Nxc5
29. Ke3 Rd7
30. Rxd7 Nxd7
31. Kd4 Kf8
32. Bh3 Ke7
33. b4 Nb6
34. Bg4 Nd7
35. a4 f5
36. exf6+ Nxf6
37. Bc8 b6
38. Ke5 Nd5
39. b5 Nb4
40. Bb7 c5
41. Be4 Kf7
42. Bb1 a6
43. bxa6 Nxa6
44. Kd6 Nb4
45. Kc7 Nd5+
46. Kc6 Nc3
47. Bc2 g5
48. Kxb6 gxf4
49. gxf4 Nd5+
50. Kb7 Nc3
51. a5
Black resigns, 1-0
* * *