The Most Instructive Positional Idea I’ve Seen in Years

Some games are memorable because of a tactic. Others because of a brilliant finish. But every now and then, a game stays with you because of a single positional idea that is so clean, so logical, and yet so unusual that it almost permanently changes the way you look at a structure.
This game between Michael Adams and Zurab Azmaiparashvili is one of those.
We begin after 11...Bb7, in a queenless middlegame arising from the Pirc with 3...e5.

This structure is very typical. White has exchanged in the centre early, which means Black no longer has to deal with the kind of broad pawn centre that often gives White a lasting edge in other Pirc lines. In return, Black has lost castling rights, and even with queens off the board, that still matters.
In the best version of these positions, Black would like to play ...c6 and tuck the king away on c7. That is usually the dream setup. But that does not always happen so smoothly, because Black still has to keep an eye on f7, and sometimes the king is nudged toward e8 instead. That is more or less what happened here.
Black has gone for a queenside fianchetto setup, while White has done everything you would want from the White side: active development, long castling, and centralized rooks. White is better developed, more centralized, and better mobilized. The pieces are on near-perfect squares. And yet, at first glance, it is not so obvious how White is supposed to turn all of that activity into something concrete. Black is passive, but solid. There are no obvious weaknesses to invade. No immediate breakthrough suggests itself.
And then Adams finds one.

The idea that makes the whole game

After 11...Bb7, White played:

12.Bxf6!

This is the move that makes the whole game worth studying.
At first sight, it looks strange. White gives up the bishop before Black has even challenged it. There is no ...h6 on the board yet. The bishop is not being asked any question. So why voluntarily part with it?
Because the real point is not the exchange itself. The point is to clear g5 for the knight.
That is the beauty of it. White understands that the bishop has done its job on g5. Now the knight needs that square more.
Once the knight lands on g5, Black is immediately forced to reckon with f7. In a normal position, Black might solve that problem by castling. But Black cannot castle here, because the queens were exchanged early and the king has already moved. That means Black is pushed into the awkward defensive move ...Rf8.
And that in turn clears the path for White’s next idea: f4, increasing the pressure on e5.
So the sequence is beautifully coherent:

• Bxf6 clears g5
• Ng5 provokes ...Rf8 and creates pressure on f7
• f4 increases the pressure on e5
• if the knight is driven back, it comes to f3, where it keeps attacking the same central weakness
• and from there White can eventually occupy d5, achieving maximum piece activity

That is why 12.Bxf6! is so deep. It is not just an exchange. It is a positional springboard.
Black replied:

12...Nxf6

Of course, taking with the pawn to stop Ng5 was never attractive. That would completely wreck the kingside structure, weaken dark and light squares alike, and leave f5 especially tender. White would then be able to reroute a knight via h4 into f5 with great effect, and Black would no longer have g6 available as a reliable way to fight for that square.
So after 12...Nxf6, White continued exactly as planned:

13.Ng5 Rf8 14.f4

And this is what I love about the sequence: every move flows logically from the previous one.
White first plays Bxf6 so the knight can come to g5.
Then the knight goes to g5 to force Black into defending f7 awkwardly.
Then White plays f4, bringing another attacker to bear on e5.
And if the knight is chased, it can simply retreat to f3, from where it still helps pile pressure onto that same central point.
This is not random activity. It is a perfectly coordinated positional plan.

e5 becomes the battlefield

Black continued with:

14...h6 15.Nf3 Nd7

This is basically forced. Black has to defend e5 somehow, but there is a price: White now gets to jump into d5, one of the most attractive squares on the board.

16.Nd5

A wonderful square. Central, active, stable, and deeply unpleasant for Black.
At this point Black understandably tries not to just sit there and get squeezed to death:

16...b5

This is the defensive side of chess in these positions. Black feels the pressure and tries to generate some counterplay on the queenside. It is like a hedgehog showing some spikes when it feels pressure coming.
It is worth noting that this is not immediately a pawn sacrifice. White cannot simply grab everything right away, because after axb5 axb5 Bxb5, Black can hit back with ...c6. So the move does have some venom.
Still, it has to be said that if Black had just played a quiet move like 16...Rd8, White’s advantage would remain, but it would not yet be so easy to convert into something concrete. White would probably continue with g3, stabilizing the kingside, and then improve the knight further with ideas like Ne3-f5 or Nh4-f5. White is clearly better, but a breakthrough is not automatic.
Also worth noting: 16...exf4? would be a serious mistake, because after e5 Black’s position starts collapsing. The bishop on d6 becomes awkwardly placed, has to keep covering c7, and has no comfortable retreat squares. Black’s whole setup begins to come apart.

The missed pawn win

Adams played:

17.Bb3

A sensible move. White does not want to open the a-file just yet.
But this was also a moment where White could have won a pawn in a very concrete way: 17.fxe5 Nxe5 18.Nxe5 Bxe5 19.axb5
And now if Black persists with the usual ...axb5 and ...c6 idea, White has the tactical shot: 19...axb5 20.Nxc7+!!
This is the key move that makes the whole variation work. If 20...Bxc7, then 21.Bxb5+ Ke7 22.Rd7+, and White wins the piece back while completely dominating the seventh rank.
That line is very instructive because it shows that Black’s queenside expansion was never fully sound. Even if Black later regains one pawn via ...Ra1 and exchanges on e1 before taking on e4, White’s connected b- and c-pawns would be extremely dangerous and the overall picture remains very grim for Black.
Still, over the board, 17.Bb3 is very understandable. It keeps the grip and avoids immediate complications.

Opening the queenside — maybe a little too soon

After:

17...Bc6 18.axb5 axb5 19.Kb1

White prophylactically sidesteps the potential ...Ra1 ideas. This is more positional than tactical: the issue is not mate, but coordination. If Black got ...Ra1 in under good circumstances, White might be forced into something awkward like Kd2, and the king would suddenly be standing on a square that interferes with the rooks rather than helping them.
Black then played:

19...Ra7

This is a strange-looking move, and I suspect it was played partly because Black was not yet psychologically ready to commit to the ugly but necessary ...f6 setup.
At first glance, ...Ra7 looks mysterious. Why lift the rook? There is no immediate need. But it may have had a few long-term ideas behind it:

• perhaps preparing to swing the rook from f8 to a8 later and create ideas like ...Ra1
• perhaps overprotecting c7 in case the bishop on d6 ever needs to move
• or perhaps just making the only move that seemed to improve the position even slightly, while Black postponed the unpleasant decision of playing ...f6

In practical terms, it feels like a waiting move: “Let me improve something and see what White shows.”
White answered with:

20.c3

This may not have been the most testing move. The idea is clear enough: White wants the option of Nb4. But compared to more direct methods, it is slightly slow and gives Black a chance to stabilize.
A more challenging route was probably something like 20.Ne3
with the idea of Nf5.
That would have posed Black more immediate problems. White should not be too worried about the central pawns being loose, because Black’s king is not comfortable enough to start grabbing material freely, and Nf5 itself is a major threat. In fact:

• exf5 is not possible because White simply recaptures with the knight and the pressure becomes unbearable
• Bxe4 would often run into fxe5, worsening Black’s king situation
• and if Black tries g6 to prevent Nf5, then Ng4 becomes very appealing

So while 20.c3 was not bad, it did not add fresh fuel to White’s attack in the same way.

Black braces with ...f6

Black now played the move they had probably wanted to avoid:

20...f6

Ugly, but necessary.
It is not the kind of move anyone enjoys making. Combined with ...h6, it weakens light squares and practically invites White’s knight to start dreaming of f5 or g6 via h4. Aesthetically, it is unpleasant.
But strategically, Black had little choice. The move finally secures e5, which had been the sore point of the entire position. And if Black were given time, they might follow up with ...g5, shutting down Nh4 ideas, or with ...Nc5, ...Bd7 or ...Kd7, after which White’s road forward becomes much less obvious.
So White now has a practical problem: if nothing active happens soon, the advantage may begin to slip.

Releasing the tension

Adams chose:

21.fxe5

This makes sense from a practical point of view. Once Black has played ...f6, White may also begin to wonder whether ...exf4 could become possible in some lines, since the old e5 break is no longer there. Even if that is not yet an immediate issue because of ideas like Nd4, White was probably encouraged to clarify the centre now.
Still, it is worth noting that White was not absolutely forced to do this immediately. There were still improving moves available, and one can argue White might have squeezed a little longer.
Black replied:

21...Nxe5

But this may have been inaccurate. Taking with a pawn would be more solid: 21...fxe5, which would at least deny White the strong Nd4 idea and bring Black much closer to equality. I can only speculate that Black was worried about White rerouting a knight to f5 or g6, but after White has already exchanged on e5, that plan is less dangerous than it looks. Black would have the f-file as a source of counterplay, could escape with the rook if needed, and might still organize ...Nc5, ...Kd7, and perhaps ...Rfa8. In that version of the game, Black is no longer just suffering.
Instead, after 21...Nxe5, White got exactly the kind of active regrouping they wanted:

22.Nd4

A very natural move. White keeps the pieces active, attacks c6 and e6 squares, and continues to squeeze.
Then came the truly strange decision:

22...Kd7?

This is the move that really lost the game.
I still do not understand why Black did not simply play 22...Bd7.
That would at least untangle the position a little. The bishop steps out of danger, e6 is covered, and Black’s pieces stop tripping over one another.
Maybe Black was worried about Nf5, but that should not have been decisive. Black could even consider taking that knight in some lines, and although White’s d5-knight might then dream of f4-e6, the resulting position is still defensible. Another idea was simply ...g6, and then White has to be careful: taking on h6 may leave the knight stranded, while Nxd6 only wins the exchange and leaves Black still in the game.
White would remain better, of course, especially with the gorgeous knight on d5, but after 22...Bd7 the position is still solid enough to fight on.
After 22...Kd7?, it is not.

The move that exposes everything

Adams now found another excellent move:

23.Nb4!

This is a superb positional shot.
White immediately senses how awkward Black’s position has become:

• the bishop on c6 is clumsy and overloaded
• it has to defend b5, and that pawn is more than just a pawn, because if White’s knight enters on b5, the position becomes nearly fatal for Black
• the bishop cannot easily move to d7, because the king is in the way
• the light squares are weak everywhere
• and White’s pieces are perfectly placed to exploit all of it

The move Nb4 increases the pressure on the c6-bishop, and with two knights now coordinating against it, Black’s problems multiply quickly.
Black cannot simply take on b4, because Nxb5+ would be crushing.
And White’s threats are nasty:

• White can take on c6 with the b-knight and then grab b5
• or White can play Be6+, after which taking on c6 wins material because the rook on a7 is hanging
• ideas like Ne6 are also in the air
• and in many fallback lines, even a simple Nf5 would leave a knight monstrously placed

The deeper point is that Black cannot solve all the problems at once.
Try to look at the position from Black’s point of view. The rook on a7 probably needs to move, because otherwise Be6+ followed by Nxc6 wins material. But the rook also has to keep an eye on b5, because that pawn cannot be abandoned lightly. Where can it go?

• Rb7 seems logical, but then White can often take on c6 twice and finish with Bd5, winning the exchange
• Ra5 is worse, because then Be6+ and Nxc6 just hit the rook directly
• Re8 would stop Be6 ideas, but it does nothing to defend b5, so White can simply win that way
• moving the bishop away, say to b7, comes too late, because after Nxb5 Black’s queenside is already falling apart

So yes, in a funny way, White did eventually benefit from having played 20.c3, because it made Nb4 possible. Even if c3 was somewhat slow at the time, it turns out to have a concrete payoff here.
There was also a very fancy way for White to win a pawn with Nxc7, grabbing a pawn defended three times. Black cannot really recapture:

• Kxc7 runs into Ne6+
• Bxc7 is a disaster after Nxb5
• and Rxc7, which is the critical recapture to calculate, runs into Nf5, when the bishop on d6 cannot be defended properly because the temporary defensive squares f7 and c4 are controlled by White’s light-squared bishop

So White would win the piece back and remain a pawn up.
Still, the move in the game was simpler and stronger. No need to be fancy when the direct route works.

Black gives up the exchange

Black chose:

23...Rfa8 24.Nbxc6 Nxc6 25.Nxc6

This wins the exchange cleanly.
There was also another equally good route to win an exchange: 25.Be6+ Ke7 26.Nxc6+ Kxe6 27.Nxa7
but Adams’ choice is simple and efficient.
After:

25...Kxc6 26.Bd5+ Kb6 27.Bxa8

White has won the exchange, and from a practical standpoint the rest is an excellent model conversion.
What I like here is that Adams does not rush. He does not allow counterplay. He improves the king, coordinates the rooks, and slowly pushes Black further and further back.
For players who sometimes struggle with conversion, it might be worth checking the rest of the game, as the conversion stage by Adams was very clean, White never gave Black even a glimpse of real activity, but I'll just skip to the end of the game:

At this point White is no longer technically up the exchange, but it does not matter in the slightest. The exchange was returned at the right moment to break through, and now the b-pawn decides the game.
Black resigned because promotion cannot be stopped in a satisfactory way. Black will eventually have to give up the rook on b7, and the checking ideas do not save them. For example, if Black tries ...Rc2+, White can step to Kb6, and if ...Rb2+, White has the classic bridge-building resource with Rb5.
A very clean finish.

Final thoughts

This game is a great example of a queenless middlegame that behaves almost like an endgame, except there are still enough pieces on the board for dynamic imbalances and long-term pressure to be a huge factor in a position.
Neither side played perfectly. Black especially made a few very strange decisions, and the move 22...Kd7 was the real turning point. But that should not distract from what White did so well.
Adams showed a very deep kind of positional understanding:

• he identified that his activity was real, even if there were no immediate entry points
• he found the unusual 12.Bxf6! idea to give that activity direction
• he followed up with complete logical consistency: Ng5, f4, Nf3, Nd5
• he kept improving his pieces until Black’s position became so strained that one inaccuracy caused the whole structure to collapse

And that is why the game is so instructive.
The move 12.Bxf6! is the part that really stays in the memory. It looks almost paradoxical: White voluntarily gives up the bishop before it is even challenged. But in reality, it is not a surrender of the bishop at all. It is an investment. White gives up one well-placed piece so that the rest of the army can reach even better squares.
Without that move, it would have been much harder for White to come up with anything meaningful. With it, the whole plan begins to flow naturally.
That is what makes it such a beautiful positional idea.