The bishop pair · For club players · 17 min read
Bishop Pair Strategy: Kramnik v Tiviakov 2001, Checked by Stockfish
Kramnik, the new world champion, turned the two bishops into a win at Wijk aan Zee 2001. Stockfish 19 shows what the pair was worth and where both players let the advantage slip.
- plan
- weakness
- target
- Where the bishops aim
- pawn chain or line of force
In short
- The bishop pair is two bishops against bishop and knight (or two knights); it grows stronger as the position opens.
- Kramnik got the pair when Tiviakov exchanged on c3, then freed the bishops with c4 and d5 and opened the f-file with f4.
- Stockfish shows the position was level at move 9; Black's small inaccuracies built a lead of about two pawns by move 19.
- Both players erred later (22.Ne4, 28.g4, 29...Ne5), and the passed pawn on d6-d7 finished the game.
How we know what we say
- Verified From the record: Stockfish's numbers, the players' own notes or interviews, or a cited source.
- Inferred Read from the moves and the position: a reasoned reconstruction, not the player's words.
- Speculative A guess about mood, intent or practical motives: plausible, not proven.
Part 1Kramnik v Tiviakov, Wijk aan Zee 2001: a world champion and the bishop pair
In January 2001 Vladimir Kramnik arrived at the Corus tournament in Wijk aan Zee as the new classical world champion: in London in 2000 Kramnik had beaten Garry Kasparov by 8½–6½ without losing a game. [4] Verified
In round 4 Kramnik, rated 2772, had White against Sergei Tiviakov, rated 2597, a solid grandmaster who had won the World Under-18 title in 1990 and who reached a peak of 2699. White accepted the pair of bishops that Black offered, and the next 38 moves show what the pair is worth and what it is not. [1][5] Verified
| Item | Detail |
|---|---|
| Game | Kramnik v Tiviakov, Corus Group A, round 4, Wijk aan Zee, 17 January 2001 |
| Ratings | Kramnik 2772, Tiviakov 2597 |
| Opening | Nimzo-Indian Defence, Rubinstein System (E59) |
| Result | 1–0 in 38 moves |
| Strategy | The bishop pair: two bishops against bishop and knight, or against two knights |
The question to hold while reading: why did Black's sound opening turn into a clearly worse position within seven moves, and what did the two bishops actually do? A book on middlegame strategy by Jan Markos uses this game as its third example of the bishop pair, so it is a recognised teaching game for the idea. [3] Verified
Part 2What is the bishop pair, and when does it matter?
The bishop pair is two bishops, one on each colour of square, against a bishop and a knight or against two knights. Together the two bishops cover every square on the board, and from far away. The knights jump short distances and need safe squares close to the action.
Chess writers often value the pair at about half a pawn, but the figure moves a great deal with the position; in this game Stockfish found it nearly worthless in the opening and worth two pawns by move 19. The pair is worth most when the position is open (few locked pawn chains), when there are pawns on both wings so the bishops have targets far apart, and when the opponent has no outpost for a knight. It is worth least in a closed, blocked position, where the bishops bite on their own pawns. Inferred
- key square
- Bishops hemmed in
- pawn chain or line of force
- key square
- Bishop's target
- pawn chain or line of force
The same two bishops in the same game. The pair is worth little while the pawns shut it in and a great deal once the lines open; Stockfish rates the right-hand position far higher for White.
Part 3The opening: Black gives up the pair on purpose
The game began 1.d4 Nf6 2.c4 e6 3.Nc3 Bb4 4.e3: the Nimzo-Indian Defence. After 8.a3 Black's bishop is attacked and Tiviakov took the knight with 8...Bxc3 9.bxc3, the standard trade in this line. It hands White the pair of bishops but leaves White with doubled, weak pawns on the c-file. Stockfish approves of Black's choice: after the exchange it still rates the position as level, within a fifth of a pawn. Verified
1.d4 Nf6 2.c4 e6 3.Nc3 Bb4 4.e3 O-O 5.Bd3 d5 6.Nf3 c5 7.O-O Nc6 8.a3 Bxc3 9.bxc3 dxc4 10.Bxc4 Qc7 11.Bb2 e5 12.h3 e4 13.Nd2 Na5 14.Ba2 Bf5 15.c4 +0.71White slightly better
The first fifteen moves. Black grabs space with ...e5-e4 and chases the White pieces, hoping the doubled pawns will matter more than the bishops. Verified
The pair is not an advantage yet. What turned it into one was Black's setup. The pawn on e4 kicked the knight from f3, but it also locked in the centre and let White prepare the break f2-f4. The knight went to a5, which attacks the bishop on c4 and sits on the rim. After 14.Ba2 Bf5 Stockfish rates White as only slightly better, by about half a pawn. Verified
- target
- outpost
- pawn chain or line of force
The doubled pawn that looked like White's weakness is the key to the whole game: it is the pawn the dark-squared bishop needs out of the way. 15.c4 costs nothing and gives that bishop its diagonal. Inferred
Part 4The plans: open the position against the stranded knight
After 15.c4 the position has a clear shape. White's bishops are about to see the whole board; Black's knight on a5 has no good squares, and Black's other pieces defend. Stockfish's early opinion tilts further with every White move: Stockfish's evaluation goes from about half a pawn after 14...Bf5 to nearly two pawns by move 19. Verified
- plan
- Target zone
- plan
- key square
- Black's grip
One position, two plans. White's bishops grow stronger with every pawn that leaves the board; Black needs to trade them off or close the position before the attack arrives.
Part 5The moments that decided it
Stockfish 19 searched every position of the game. The graph shows a steady climb for White through move 19, a drop at move 22 when White trades a knight for a rook, and the large swings in the last third.
Critical moment 1 of 4 · Move 16, White to play
How do the bishops get their lines?
- weakness
- key square
White has the pair; Black's knight is stuck on a5 and the centre is half closed. Which pawn move gives the bishops targets and the knights nowhere to go?
Think it through on the boardWeigh the candidates
| Move | Stockfish | Verdict |
|---|---|---|
| 16.Qe2 16...Bg6 17.d5 b6 18.Bb1 | +0.86White slightly better | Close to the game move: Black's bishop retreats and White still gets d5, with an edge of a bit under a pawn. A good alternative. Verified |
| 16.Qc2 16...cxd4 17.Bxd4 Rad8 18.f4 | +0.60White slightly better | Invites ...cxd4 and an exchange of pawns that frees Black's pieces; Stockfish gives White about half a pawn, well behind the game move. Verified |
| 16.Re1 16...Nc6 17.d5 Ne5 18.Bb1 | +0.55White slightly better | Lets the knight come back via c6 and e5; about half a pawn for White and the weakest of the three. Verified |
What Kramnik played
16.d5+0.97White clearly better
Kramnik pushed the d-pawn to d5. It grabs space, takes c6 and e6 from Black's knights and leaves the knight on a5 with no way home.
Stockfish's first choice too. Depth 25.
What it revealed
Opening the position is not always done by exchanging pawns. A pawn that restricts the enemy knights is as good as one that opens a file, and here it also created the passed pawn that decided the game fourteen moves later.
- Verified Stockfish agrees with 16.d5: it is the engine's first choice, a little ahead of 16.Qe2 and a third to half a pawn ahead of the other quiet moves.
- Inferred The pawn on d5 also gives White a future passed pawn on the d-file, which is exactly what decides the game.
Black's reply 16...Nd7 and the move before it, 15...Rfe8, were the real concessions. Stockfish preferred the bishop retreat ...Bg6 both times. Counting every Black move from 12...e4 to 16...Nd7, Stockfish finds five small losses of 12, 8, 20, 37 and 65 hundredths of a pawn: none of them an obvious blunder, together about one and a half pawns. Verified
Critical moment 2 of 4 · Move 22, White to play
Keep the bishops, or cash them in?
- target
- Bishop's diagonal
White is clearly better and both bishops are active. The knight on d2 can go to e4 and be exchanged for a rook. Is it better to keep every piece and the pressure, or to turn it into material?
Think it through on the boardWeigh the candidates
| Move | Stockfish | Verdict |
|---|---|---|
| 22.Bf5 22...Ne5 23.Qf4 Bxf5 24.Qxf5 | +1.84White clearly better | Stockfish's choice by a wide margin: the bishop challenges the one on g6 and White keeps every piece and the whole bind, rated at nearly two pawns. Verified |
| 22.Qh3 22...Ne5 23.Bxg6 fxg6 24.Bxa5 | +0.91White clearly better | Aims at the bishop and the h-file; about nine-tenths of a pawn, a little better than the game move. Verified |
| 22.Qf2 22...Re7 23.Bf5 Rae8 24.e4 | +0.27about equal | Too slow: Black gets time to untangle and the edge falls to about a quarter of a pawn. Verified |
What Kramnik played
22.Ne4+0.70White slightly better
Kramnik played 22.Ne4. The knight is offered to the rook: after 22...Rxe4 23.Bxe4 White wins the exchange (a rook for a knight), and the bishop on e4 now sits on the b1-h7 diagonal.
Stockfish preferred 22.Bf5 (+1.73); the game move gave up about 1.0 pawns. Depth 28.
What it revealed
Kramnik chose to turn positional pressure into material. It is a natural instinct for a player who is better and wants a simpler game, and it works as an exchange for a pawn after 23...Nxc4. But it lets Black out of the bind.
What Stockfish adds Stockfish prefers 22.Bf5 and puts the game move about a pawn behind it. The exchange is real, but Black wins a pawn back at once with 23...Nxc4 and the position loosens.
- Verified Stockfish rates 22.Bf5 about a pawn higher than the game move: the evaluation drops from nearly two pawns to well under one.
- Inferred Winning the exchange is concrete and easy to explain, which is probably why it is tempting; keeping the bind needed the harder, quieter bishop move.
Critical moment 3 of 4 · Move 24, Black to play
Should Black trade queens?
- key square
- target
White has won the exchange for a pawn and holds two bishops. The queens face each other on d6 and f4. Does Black want them off?
Think it through on the boardWeigh the candidates
| Move | Stockfish | Verdict |
|---|---|---|
| 24...Nde5 25.Bxg6 fxg6 26.Rac1 Re8 | +0.34about equal | Stockfish's choice: keeps the queens on and White's edge falls to about a third of a pawn. Verified |
| 24...Nce5 25.Bf5 Bxf5 26.Qxf5 g6 | +0.50White slightly better | Nearly as good: White keeps about half a pawn. Verified |
| 24...b5 25.a4 Qxf4 26.Rxf4 Bxe4 | +2.98White winning | Loses at once: White is close to three pawns ahead. Verified |
What Tiviakov played
24...Qxf4+1.53White clearly better
Tiviakov took the queen with 24...Qxf4 25.Rxf4. It looks natural: Black cuts down White's attack and keeps the knights active.
Stockfish preferred 24...Nde5 (+0.31); the game move gave up about 1.2 pawns. Depth 26.
What it revealed
The queen trade helps the side with the bishops. Without queens nothing can harass the bishops, and the open board is what they need. A knight move that kept the queens on would have kept White's edge small.
What Stockfish adds After 24.Qf4 the engine had White only a third of a pawn ahead, because Kramnik's 24th move was itself half a pawn short of 24.Bxg6. The queen trade gave the edge back and more.
- Verified Stockfish says 24...Qxf4 gives away more than a pawn compared with 24...Nde5.
- Inferred Black's instinct, to trade queens when the opponent has the initiative, is usually sound; here it just trades the one piece that could harass the bishops.
Part 6The turning point: 29...Ne5
By move 27 Tiviakov had grabbed two pawns with the knights and material was level, but White's two bishops, the d5 pawn and a stranded knight on c2 gave Stockfish a lead of over two pawns. Kramnik then played 28.g4, which gave much of it back. Stockfish's choice was 28.Bxh5!, which pulls the bishop on g6 away from the defence of the knight on c2: after 28...Bxh5 29.Rxc2 White wins a piece. Tiviakov, given a second chance, replied with the move that decided the game. Verified
Critical moment 4 of 4 · Move 29, Black to play
Black to move: the rescue that was not
- target
White's bishop on g4 is guarded only by the rook, the pawn stands on d5, and Black's knight on c2 is cut off. Which move holds the game together best?
Think it through on the boardWeigh the candidates
| Move | Stockfish | Verdict |
|---|---|---|
| 29...Rd8 30.h5 Bd3 31.h6 gxh6 | +0.72White slightly better | Stockfish's choice and the only move that keeps White's edge under a pawn and a half: White stays about two-thirds of a pawn ahead. Verified |
| 29...Nf8 30.Bf5 Bxf5 31.Rxf5 Nd4 | +2.14White winning | Passive: White is more than two pawns ahead. Verified |
| 29...Nf6 30.Bxf6 gxf6 31.d6 Rd8 | +4.78White completely winning | Wrecks the kingside pawns and lets d6 in: nearly five pawns for White. Verified |
What Tiviakov played
29...Ne5+2.42White winning
Tiviakov played 29...Ne5, attacking the bishop on g4. The knight is protected and the capture on g4 looks like it breaks the pair.
Stockfish preferred 29...Rd8 (+0.66); the game move gave up about 1.8 pawns. Depth 25.
What it revealed
Black went after the bishops, a natural reaction to the pair, but the knight left the d-file and the defence of the d-pawn's path. 30.d6 followed, and once the knight took on g4 the pawn on d6 was decisive.
What Stockfish adds 29...Ne5 gives away about a pawn and three-quarters, and 30...Nxg4 gives away nearly as much again. The engine's defence is to put the rook on the d-file and leave the bishops alone.
- Verified Stockfish gives 29...Rd8 as the only move that keeps White's edge under a pawn; the game move 29...Ne5 swings the evaluation by about a pawn and three-quarters.
- Inferred The idea of trading a bishop is right in principle, but the passed d-pawn was worth more than the pair itself.
- Speculative Tiviakov may have been drawn by the textbook advice to trade off one of the bishops when facing the pair; here the trade cost the game.
- plan
- key square
- target
- pawn chain or line of force
33...Rd6 34.Rxg7+ Kf8 35.h6 Nd4 36.Rag2 Bg6 37.Rf2 Nf5 38.h7 +7.20White completely winning
The finish: the rook takes on g7 with check, the h-pawn marches to h7, and the game ended there with White about seven pawns ahead by Stockfish's count. Verified
Part 7What Tiviakov could have done
Black was worse for most of the game, but the engine finds a stubborn defence at move 29: 29...Rd8. The rook goes to the d-file and stops the pawn at once. Stockfish keeps White ahead by about two-thirds of a pawn, which is something but far less than the three and a half pawns of the game. Verified
- outpost
- target
29...Rd8 30.h5 Bd3 31.h6 gxh6 32.Bh5 f6 33.d6 +0.49White slightly better
Stockfish's defence: Black blockades with the rook and the game goes on; White presses with h5 and the bishops but the pawn on d5 is held. Verified
Part 8A fresh look: five small concessions, not one big mistake
The usual story of this game is simple: the two bishops swept Black off the board. Stockfish tells a different story. In the opening, level at move 9, the pair was worth almost nothing. The advantage arrived through a run of small Black inaccuracies (12...e4, 13...Na5, 14...Bf5, 15...Rfe8 and 16...Nd7), each one a fraction of a pawn, together worth about one and a half pawns. The pair did not create the lead; it made every small error expensive. Verified
| Move | What happened | Evaluation after (White's view) |
|---|---|---|
| 8...Bxc3 | Black gives White the pair | about +0.2 |
| 14...Bf5 | Black's setup is slow | about +0.6 |
| 15...Rfe8 | misses 15...Bg6 | about +1.1 |
| 16...Nd7 | misses 16...Bg6 | about +1.6 |
| 19.h4 | White's bishops press on | about +2.0 |
| 22.Ne4 | White trades a knight for a rook; 22.Bf5 kept nearly +1.7 | about +0.7 |
| 24...Qxf4 | Black trades queens; 24...Nde5 held at +0.3 | about +1.5 |
| 26...Nc2 | the knight is cut off | about +2.0 |
| 28.g4 | White gives much of it back | about +0.8 |
| 29...Ne5 | Black gives it all back | about +2.4 |
| 30...Nxg4 | the pawn reaches d6 | about +3.6 |
Both players erred in the last third. Kramnik's 28.g4 threw away about a pawn and a half, and 22.Ne4 about a pawn. A player who owns the pair does not have to be perfect, only to keep the position open and the opponent's pieces busy. The pair keeps giving second chances. Inferred
The decision at move 22, to win the exchange instead of keeping the bind, is the mirror image of an idea covered in the exchange sacrifice article, where the exchange is given up for lasting pressure; compare also bishop against knight.
Part 9How to play with the bishop pair in your own games
The recipe below works for club players because it does not need engine precision: it asks what the pieces want and removes what stops them. Inferred
The recipe for the pair
- Count the piecesBefore you trade a bishop for a knight, see whether you would be giving up the pair. If you can keep both bishops at no cost, do.
- Clear the bishops' own pawnsLook at each bishop and name the pawn that blocks it. A move like 15.c4 can open a diagonal without any exchange.
- Open the position where the king or weaknesses areUse a pawn break (f2-f4) to open the side the bishops aim at, not the one where the knights are happy.
- Take away the knight's squaresPlace pawns so the knights have no outposts. A pawn on d5 took c6 and e6 from Black's knights.
- Cash in when it is concreteIf you can win material or a passed pawn, take it. A tidy exchange is better than an attack that needs everything to work.
- Do I still have both bishops?
- Which of my pawns blocks a bishop?
- Which break opens the position where I am strong?
- Where can the enemy knight land safely?
- Can I turn the pair into something concrete: a passed pawn, the exchange?
Your turnPractise it
Set each position up, decide your plan before you look, then play it out against Stockfish to see whether it holds.
White to play (move 16): Black's knight is stuck on a5. Which pawn move cramps Black and prepares to open the position for the bishops?
Hint
Look at which squares the black knights want, c6 and e6, and which pawn can take them away.
A good answer
16.d5, Stockfish's choice. It takes c6 and e6 away from the knights and prepares f4. The quiet queen moves are weaker by a quarter to half a pawn.
White to play (move 28): Black's knight on c2 is cut off from the game. Is there a move that wins it, or must White give something back?
Hint
Which black piece guards the knight on c2, and can it be deflected?
A good answer
28.Bxh5, Stockfish's choice: the bishop on g6 guards the knight along the diagonal g6-c2, so if it recaptures on h5 the rook takes the knight on c2. The game move 28.g4 let Black off, and the evaluation fell by a pawn and a half.
Black to play (move 29): White threatens a pawn march on d6. Which move limits the damage?
Hint
Which file does the pawn need, and which piece can stand in front of it?
A good answer
29...Rd8, putting the rook in front of the pawn. The game move 29...Ne5 allowed 30.d6.
Part 10What to take from this game
- The pair needs open lines. Open the position where the enemy king or weak pawns are, and keep the bishops pointed at them.
- Small concessions add up. Five inaccuracies of a fraction of a pawn each became a clearly worse game; the pair made each of them count.
- Clear your own pawns. A bishop blocked by its own pawn can be worth less than a knight: 15.c4 freed the diagonal.
- Defend by blockading. The passed pawn was more dangerous than the bishops; 29...Rd8 was the defence.
- Cash in the advantage. The pair is a means, not a result: turn it into a passed pawn or the exchange.
Questions readers ask
What is the bishop pair in chess?
The bishop pair is having both of your bishops while the opponent has only one bishop, or none. Having bishops on both colours of square lets them cover the whole board, which is why they are strong in open positions.
How much is the bishop pair worth?
It is often valued at about half a pawn, but the value changes with the position. In an open position with pawns on both wings it can be worth much more; in a closed position it can be worth nothing.
When is the bishop pair not an advantage?
When the position is closed. If the pawns are locked into chains, the bishops bite on their own pawns and knights with outposts are as good or better. Look for a break that opens the position, or give up the pair for a concrete gain.
Should I trade a bishop for a knight?
Only for something concrete: structure, a weak pawn or the initiative. Giving up the pair for nothing hands the opponent a lasting edge in open positions, as Tiviakov found against Kramnik in 2001.
How do I play against the bishop pair?
Close the position or trade one of the bishops, put your knights on outposts the bishops cannot challenge, and blockade passed pawns. In this game Black's best chances lay in 29...Rd8, which blockaded the d-pawn.
Sources and method
The game score was checked against two independent sources. Every position, line and candidate in this article was replayed for legality, and every engine number was produced by Stockfish 19 WASM Multithreaded for this article (depth 13 to 32, deepest at the critical moments), not copied from elsewhere. What the players thought is reconstructed from the moves, their published notes and later analysis; it is labelled as verified, inferred or speculative, and nothing is presented as a quotation unless a source is cited for it.
- Kramnik v Tiviakov, Corus Group A, Wijk aan Zee 2001: game score and PGN (chessgames.com)
- Vladimir Kramnik, games archive (PGN Mentor), the same game score
- Jan Markos, Middlegame Secrets Vol. 3: The Career Paths of Bishops (contents list: The Bishop Pair 3, Kramnik v Tiviakov)
- Vladimir Kramnik (Wikipedia)
- Sergei Tiviakov (Wikipedia)
Read next
In game 7 of the 1971 Candidates final, Fischer traded a dominant knight for a bad bishop and won a textbook bishop-against-knight ending. Stockfish 19 says the trade was not best. Here is why it worked anyway.
Outposts: Kasparov's Octopus Knight v Karpov, Checked by StockfishKasparov's knight on d3 in game 16 of the 1985 match, what Stockfish says about it, and how to build an outpost no pawn can attack.
King in the Centre: Morphy's Opera Game, Checked by StockfishMorphy's famous 1858 win against two amateurs, and what Stockfish 19 says about when the uncastled king was really lost.
Turn reading into skill
Strategy sticks when you use it. Practise real positions with Stockfish's verdict at once, or play the plans out from any diagram above.