Xiangqi Strategy Guide: Openings, Tactics & Endgame
Xiangqi hides nothing and shuffles nothing, so every result comes from the quality of your plans. This page takes you from opening priorities to mating patterns and endgame technique.
These are practical principles, not a solved opening book. Because Xiangqi is a perfect-information game — no hidden setup, no dice — better decisions win games, and the decisions below are the ones that matter most at every stage. Everything here follows the ruleset Paofuqi actually enforces, including the two endings beginners misjudge most often: stalemate is a loss, not a draw, and a game runs to a draw after 100 half-moves with no capture. If you still need the rules themselves, start with the complete Xiangqi guide.
Step 1: Open with development, not memorized traps
Red moves first, and every game begins from the identical position — so unlike Banqi, opening theory in Xiangqi is real and deep. Memorizing move orders is still the slowest way to benefit from it. What wins games below master level is a short list of priorities you can apply to a position you have never seen.
1. Bring the strong pieces out early. Chariots, Horses and Cannons do nearly all the fighting. A Chariot still on its home point behind an undeveloped Horse is worth a fraction of its value — the most common reason club players lose is simply that one side finished developing first.
2. Contest the central file. The middle file points straight at the enemy palace, which is why the central cannon (炮二平五 — swinging a Cannon to the middle) is the most played first move in the game. It creates an immediate threat down the file and takes the most valuable line on the board.
3. Answer a central cannon before you counter-attack. The standard reply is the screen horses defence (屏風馬): both Horses step up to cover the central Soldier and each other. The diagram below shows this exact skeleton — the shape behind a large share of all Xiangqi games.
4. Do not push rim Soldiers without a reason. Advancing the Soldiers on files 1 and 9 gains almost nothing early and permanently loosens the squares around your Horses.
- Ask what the opponent's last move threatens — in a perfect-information game the threat is always visible.
- Develop a piece that is doing nothing, preferring Chariots to open files.
- Check whether the central file is safe for you and dangerous for them.
- Only then start a concrete attack.
Step 2: Know what each piece is worth
Xiangqi trades are decided by value, and the values are not the ones chess players expect. The table below is the exact material scale used by the Paofuqi computer opponent, expressed in Soldiers so it is easy to hold in your head.
Two adjustments matter more than the raw numbers. A Soldier doubles in value once it crosses the river, because only then can it move sideways — a crossed Soldier sitting next to the enemy palace is a genuine attacker, not a pawn. And Advisors and Elephants are defenders that cannot cross to help you attack: they are worth two Soldiers while your palace is under fire, and much less when it is not.
| Piece | Worth (in Soldiers) | What decides its real value |
|---|---|---|
| Chariot 車 | 9 | Open files. A Chariot on a blocked file is doing nothing; on an open one it is the strongest piece on the board. |
| Cannon 炮 | 4.5 | Screens. Cannons are strongest early, when the board is crowded with pieces to jump over, and weakest in an empty endgame. |
| Horse 馬 | 4 | Space. Horses are weak in the cramped opening because the leg is easily blocked, and grow stronger as the board opens — the mirror image of the Cannon. |
| Advisor 仕 / Elephant 相 | 2 each | Your own palace. Pure defenders; they can never cross the river to join an attack. |
| Soldier 兵 | 1 (about 2 across the river) | The river. Before it, forward only; after it, forward or sideways — a completely different piece. |
The practical consequence: do not trade a Cannon for a Horse early just because the numbers look close — the Cannon is at its peak while the board is full. In the endgame the same trade is usually good for you.
The other consequence is that a Chariot is worth roughly two Cannons or two Horses. Losing a Chariot for a lesser piece is the single most common way a winning position turns into a lost one.
Step 3: Learn the mating patterns
Xiangqi mates are pattern-based, and two of them account for a large share of all decisive games. Both exploit the same weakness: the General is trapped in a 3×3 palace, so it runs out of squares faster than a chess king ever does.
The smothered mate (悶宮) uses the palace against its owner. A Cannon fires down a file at the General while the General's own Advisors seal its sideways escapes and one more friendly piece blocks the square in front — the same piece that serves as the Cannon's screen.
The horse-and-cannon mate (馬後炮) is the other pattern worth knowing by heart. A Horse sits directly in front of the General, covering both diagonal escape points, while a Cannon behind it uses the Horse as its screen to attack the General itself.
Step 4: Avoid the five common mistakes
1. Forgetting that your General's file is a weapon — and a liability. The two Generals may never face each other down an open file. That rule does more than end games: any piece of yours standing alone between the two Generals is pinned to that file, because stepping off it would create the illegal position. The engine simply never offers you those moves, which is exactly when players get confused about why a piece appears stuck.
2. Breaking up your Advisors and Elephants. They are the only defenders your palace will ever have, and they cannot be replaced. Trading them off to win a Soldier hands the opponent the smothered-mate and horse-and-cannon patterns from the previous step.
3. Grabbing material with a Chariot. A Chariot that wanders off to capture a Soldier is usually one move away from being trapped by a Horse and a Cannon. Ask what the Chariot is doing after the capture, not just what it takes.
4. Treating stalemate as a draw. In the Paofuqi ruleset — following the WXF and Chinese convention — a player with no legal move loses. This is the opposite of Western chess, and it changes endgames completely: sealing the enemy General with no escape is a win, not a save.
5. Shuffling in a drawn-looking endgame. A game with no capture for 100 half-moves (50 moves each) is declared a draw. If you are the one trying to win, that clock is running against you; make progress or force a capture.
Step 5: Convert the endgame
Endgames are where Xiangqi's values invert. Cannons lose power as screens disappear, Horses gain it as space opens, and Soldiers that crossed the river become real attackers. Three techniques convert most winning positions.
Cut the palace with a Chariot. Placing a Chariot on the rank directly in front of the enemy palace confines the General to its back rank without giving check. From there you bring up your own pieces while the enemy General has nowhere to go.
Use your own General. In the endgame the General is an active piece. Bringing it up the central file removes squares from the enemy General through the facing rule, and it is often the move that turns a stalled position into a mate.
Push crossed Soldiers together. A single Soldier rarely finishes a game, but two Soldiers across the river, supported by a Chariot or your General, are a genuine mating force. Advance them side by side so they defend each other's squares.
Finally, count the no-capture clock. If you are winning but the position has been quiet for a long time, prefer a line that forces a capture or a decisive threat over one that merely improves your position — the 100-half-move draw does not care who is better.
Play Banqi free