Caro-Kann Defense: Endgame Offer: Bg4 – Playing White with Confidence
After 1.e4 c6 2.Nf3 d5 3.d3 Bg4, Black pins your knight. You push them off it with 4.h3 — but what comes next? You're playing White, and the engine gives +0.40, a small but clear edge in your favour. That means you stand slightly better right now. Below you'll play the position against an adapting engine: the drill will teach you the key trade on f3, the most common continuations, and the mistakes Black makes most often. Let's turn your slight edge into a real plus.
Play the Caro-Kann Defense: Endgame Offer: Bg4 against the engine
Free, no signup — you play white, the engine adapts to your level.
Play the position now against an adapting engine. The drill will show you the key moments — trade on f3, punish inaccuracies, and turn your +0.40 edge into awin
Create a free account →The Big Idea: Offer to Trade on f3
Black's bishop on g4 is pinning your knight against your queen. With 4.h3 you ask: trade or retreat? The engine's best answer for Black is Bxf3, giving up the bishop pair right away. After 5.Qxf3, Black usually plays 5...e6, and you continue with Nd2. You'll have the bishop pair, a solid centre, and easy development. The statistics back this up — across over 47,000 games where Black took on f3, White scores 43.5%. That's healthy in an opening where Black wins overall 53.2% of the time from this position. Your plan: finish developing, control the centre with c3 and e4, and enjoy having the two bishops in a quiet game.
What Black Actually Plays (and What It Means)
From this position, Black chooses between two main paths, and their choice changes your approach entirely. Let's look at the numbers from 100,311 games at this exact position (White wins 43.0%, draws 3.8%, Black wins 53.2%):- Bxf3 (47,969 games): This is the principled trade. White scores 43.5% — solid. You trade queenside development for the bishop pair. After 5.Qxf3 e6 6.Nd2 you're slightly better, with a stable centre and long-term pressure.- Bh5 (46,556 games): Almost as popular as the trade. Black retreats, keeping the pin alive. Here White scores 42.0% — still respectable. You can continue with 4...Bh5 5.exd5 cxd5 6.g4 Bg6 7.Ne5, gaining space and tempo on the bishop.Neither option is scary. The engine prefers Bxf3, but both are playable for both sides.
Punish the Inaccuracies: dxe4, Be6, and Bd7
The engine flags three Black moves as clear inaccuracies — and knowing them means you can spot when your opponent goes wrong. If Black plays dxe4, they lose roughly 0.8 pawns of advantage compared to Bxf3. The correct reply is simply 5.dxe4, opening the centre while Black's bishop is misplaced. Similarly, Be6 and Bd7 are both inaccuracies costing about 0.7 pawns each. Against Be6, you can play 5.exd5 Bxd5 6.c4, chasing the bishop and gaining space. Against Bd7, you simply continue with normal development — Black has wasted a tempo moving a bishop twice while you advanced in the centre. The lesson: if Black avoids the trade on f3, they need to pick Bh5, not any of these.
The Typical Middlegame You're Aiming For
After the main line (4.h3 Bxf3 5.Qxf3 e6 6.Nd2), you have a comfortable position with clear plans. Your bishop pair is your long-term asset — White's light-squared bishop is active, and the dark-squared one will find a good diagonal. Your knight on d2 eyes c4 and f3 squares. Black usually plays 6...Nf6, and you can continue 7.e5 Nfd7 8.Bd3, clamping down on the centre. You'll castle kingside, play c3 to support d4, and look to slowly squeeze Black. The pawn structure is closed but not locked — you have space on the kingside, and your bishop pair can exploit any weakness Black creates. It's a patient, positional game where your slight opening edge (engine says +0.40 in your favour) can grow if you keep the pressure.
Results across 100,311 Lichess games
| Most-played continuation | Games | White wins |
|---|---|---|
| Bxf3 | 47,969 | 43.5% |
| Bh5 | 46,556 | 42.0% |
| dxe4 | 5,122 | 45.7% |
| Be6 | 237 | 54.0% |
| Bd7 | 165 | 52.1% |
| e6 | 123 | 71.5% |
Frequently asked questions
Is 4.h3 the best move in the Caro-Kann Endgame Offer Bg4?
Yes — 4.h3 is White's most challenging reply, asking the bishop to reveal its intentions. The engine gives +0.40, a small edge for White, meaning you stand slightly better after this move. Other fourth moves are possible, but h3 forces Black to either trade on f3 or retreat, both of which are comfortable for White.
What should White do after Black plays Bh5 on move 4?
Bh5 is Black's second most popular reply (46,556 games). Your plan is straightforward: play 5.exd5 cxd5 6.g4 Bg6 7.Ne5. You gain space on the kingside and centralise your knight. White scores 42.0% from here — perfectly playable. Just keep developing and don't overextend.
Why is dxe4 a mistake for Black in this position?
The engine calls dxe4 an inaccuracy, losing about 0.8 pawns compared to Bxf3. You simply recapture 5.dxe4, and Black's bishop on g4 is awkward — it can't take on f3 without cost, and it's vulnerable to h3 ideas. White gets a comfortable centre with Black's bishop poorly placed.
How can White improve their score from 43.0% wins in this position?
Focus on the main line: when Black trades on f3 (which they do in roughly 48% of games), play 5.Qxf3 e6 6.Nd2 and build a strong centre with c3, e5, and Bd3. Avoid unnecessary complications — this is a positional edge, not a tactical crush. Patience and the bishop pair will do the work.