4 vs 3 Rook Endgame
The rook endgame with four pawns against three on the same wing is the ending grandmasters reach every single week: someone won a queenside pawn in the middlegame, everything else got traded, and the extra pawn turns out to live on the crowded kingside where it is hardest to convert. With accurate defense the 4 vs 3 on one wing is a draw — that has been the classical verdict for a century and modern engines confirm it — but the draw must be earned over decades of moves, and most club players lose it. This page gives the defensive setup, the pressing method, and the tablebase-verified endpoints the whole struggle funnels into.
The Tabiya: What Both Sides Are Playing For
Here is the structure in its typical form — White pressing with pawns e4, f3, g3, h4 against f7, g6, h5, both rooks active:
Look at Black's arrangement, because it is the recommended one: pawns on f7, g6 and h5, king on g7, rook behind White's position on the second rank. The pawn chain concedes nothing: f7 stays home so the king keeps shelter, g6 blunts the e4–e5–e6 spearhead's support points, and the advanced h5-pawn is the single most useful defensive move in the whole ending — it fixes White's h-pawn, volunteers itself for a trade the moment White plays g3–g4, and denies White the slow g4/h5 space-gaining clamp. Each pawn trade is a defensive victory: fewer pawns means fewer levers, and the endpoints below explain why the defender wants precisely the f- and h-pawns to be the survivors on White's side.
The active rook is not negotiable. A passive rook lets White improve everything for free, then choose the moment to break; a rook harvesting tempi against White's king and pawns forces White to burn moves on defense. The general law that activity outranks material — the whole theme of rook endgame principles — applies here at maximum strength.
The Defender's Rules
- Play ...h5 early. Before White clamps with g4 and h5 himself. After a later g4, capture — ...hxg4 fxg4 leaves White's structure looser and one trade closer to the draw.
- King stays on g7. It guards f7 and g6, watches both entry squares on the f- and h-files, and never strays to where a rook check costs a pawn.
- Rook maximally active, ideally on the second rank or harassing from behind; concede a pawn for real activity rather than defend it passively. Many tablebase draws in this family are a pawn down with an active rook.
- Trade pawns, not rooks — and welcome the right liquidations. Rook trades generally help whoever's pawn ending is winning, so calculate any offer against king and pawn theory before accepting. Pawn trades, by contrast, almost always run downhill toward the drawn two-pawn endings.
- Do not create targets. Every defensive pawn move beyond the setup above is a potential hook for White's levers; when in doubt, move the rook.
The Pressing Method
The winning attempts all follow the same script, and knowing it helps both sides. White centralizes king and rook first — no pawn moves until every piece stands optimally, in the spirit of do-not-hurry technique. Then comes e4–e5, gaining space and threatening e6; then the g- and h-pawns advance to pry open a file with a lever — h4–h5, or g4–g5 against ...h5 — aiming to create a single passed pawn and an entry road for the king simultaneously, the classical two weaknesses squeeze compressed into one wing.
The defender's task is triage: meet the levers with exchanges, keep the rook annoying enough that White's king can never fully commit to escorting the passer, and steer every simplification toward the drawn endpoints. The presser's practical hope is not a forced win — there isn't one — but the accumulation of small problems: one imprecise pawn move, one passive rook retreat, one mistimed trade, and the defense collapses into a lost single-pawn ending.
The Endpoints: Where the Struggle Actually Resolves
Beneath the maneuvering, the 4 vs 3 is a negotiation about which smaller ending it becomes, and those endings are tablebase-charted. Three verdicts drive every decision:
The f+h draw. If the defender trades the g-pawns and one center pair, the remaining rook and two pawns versus rook with f- and h-pawns is a theoretical draw — the defender's promised land. King to g7, rook to long range, capture the h-pawn when it advances:
1...Rb2 waits, and when the h-pawn lunges with 2.h6+, Black takes it — 2...Kxh6! — and holds the single f-pawn ending with standard technique. Every pawn trade the defender achieves in the 4 vs 3 is a step toward this diagram.
The connected-passers loss. If White ever converts the wing into two mobile connected passers with rook support — say e- and f-pawns with the defender's king mistimed — the ending is a forced win. This is what the defender's ...h5 and lever-meeting exchanges exist to prevent: never let the pawn mass roll forward as a phalanx.
The single-pawn map. If everything trades down to one white pawn, the result follows the rook and pawn versus rook map — and here the defender's g7-king placement pays its final dividend. Even the dangerous f-pawn on its sixth rank is a verified draw when the defending king reaches the short side in time, holding with the short-side defense: king to g8, rook checking from the a-file, and the promotion square covered at the critical moment.
Read the three verdicts together and the defender's whole strategy writes itself: trade toward f+h or a lone pawn, keep the king where those defenses need it, and never allow connected passers. The attacker's strategy is the negative image: preserve the g-pawn, seek the phalanx, and drag the defending king away from g7 before cashing anything in.
Frequently Asked Questions
Is the 4 vs 3 rook endgame on one wing a draw?
Yes, with accurate defense — this is the century-old classical verdict and engine analysis supports it. In practice it is one of the most frequently lost "drawn" endings in chess, because the defense demands precision for a very long time.
What is the correct defensive setup?
Pawns on f7, g6 and h5, king on g7, rook as active as possible — ideally behind the enemy position. The early ...h5 fixes White's kingside and prepares to trade a pair of pawns the moment White plays g4.
Why does the defender want to trade pawns?
Because the reduced endings are tablebase draws: rook plus f- and h-pawns versus rook is drawn with the king on g7, and most single-pawn endings are drawn with the defending king placed correctly. Fewer pawns also means fewer levers for the attacker's breakthrough.
How does the stronger side try to win?
Centralize king and rook fully, advance e4–e5 for space, then use the g- and h-pawns as levers to create a passed pawn and an entry path for the king at the same time. The wins come from accumulated small defensive errors, not from a forced sequence.
Should the defender ever trade rooks?
Only after verifying the resulting pawn ending is drawn, which usually requires the king to reach the key blockade squares in time. As a rule of thumb the defender trades pawns eagerly and rooks reluctantly.
Why is ...h5 such an important move?
It stops White from gaining space with g4 and h5, fixes the white h-pawn as a future target, and guarantees a pawn trade when White eventually plays g4. All three effects push the position toward the drawn reduced endings.