The Counter-Palace Horse

Defence by prevention

The signature move is the cannon sliding to the fourth file - the advisor corner - early, before black’s second horse develops. On that square the cannon looks passive and is anything but: it covers the river squares red’s horse needs for the standard invasion routes, pre-empts the cannon-trading tricks red uses against the pure screen setup, and quietly supports a later central counterthrust. Red’s attacking manual assumes certain squares will be available. The fangongma’s opening bid is: they will not be.

The cost is real and worth naming: the cannon on the advisor-corner file no longer defends along black’s back rank, and black’s queenside can end up a shade slower than in the Screen Horses. The system is a wager that what red loses in attacking schemes outweighs what black pays in coordination - a wager professional practice has backed for half a century, provided black knows why each piece stands where it does.

Play it once you already understand the Screen Horses. The two defences share a skeleton, and the fangongma’s ideas only make sense as refinements of the older wall.

Hu Ronghua’s laboratory

The Counter-Palace Horse existed in the old literature as a minor sideline, and it stayed minor until Hu Ronghua picked it up. The dominant champion of the 1960s and 70s - and the most inventive opening mind of his era - rebuilt the system move by move, used it against the best Central Cannon players in China, and literally wrote the book on it. His games turned an eccentric setup into respectable main-line theory.

In our archive’s era the system remained a fixture of top practice - a defence you bring out when the opponent has prepared deeply against the Screen Horses and you would rather they spend the evening in your theory than theirs. Its games have a recognisable temperament: fewer fireworks, more slow squeezes, and a striking number of red attacks that simply never manage to begin.

The model line, move by move

Watch black’s move order carefully - it is the system’s fingerprint. The queenside horse comes first, the cannon slips to the advisor corner before the kingside horse commits, and only then does the familiar wall assemble. By step ten the position resembles a Screen Horses game in which several of red’s best ideas have quietly ceased to exist.

The line ends with a chariot trade deep in black’s camp - red’s most common practical try, cashing the initiative for a small structural pull before black’s setup reaches full strength.

The ideas underneath

The advisor-corner cannon is a square-control investment. List what it watches: the river points red’s horse needs for the classical invasion, the lateral fourth-file lanes red’s own manoeuvres cross, and - after black’s central soldier advances - the long central diagonal of counterplay. None of those jobs is glamorous; together they amount to red’s attack arriving two tempi late everywhere it tries to land.

Flexibility is the second asset. Because the cannon committed early and the horses stayed home longer, black keeps a menu red cannot read: which soldier advances, which side the chariots favour, whether the counterthrust comes in the centre or on the wing. Fangongma players win a lot of games simply by being harder to prepare against.

And the defence has a temper: when red overextends trying to make something happen against the prepared squares, the same quiet cannon becomes the counterattack’s rear battery. The system’s wins often feature that exact reversal - prevention flipping into punishment in a single move.

Plans for both sides

Black’s plan comes in phases: complete the shape, blunt red’s first wave (often by inviting exactly the chariot trade the model line shows), then choose the counterplay the position offers - the central soldier thrust supported by the corner cannon is thematic, and the re-routed kingside horse usually leads it. Black is rarely in a hurry; the system’s whole economy runs on red’s impatience.

Red’s respectable tries: the immediate central clarification the model line ends with, playing for a small enduring pull; a slow build that refuses to touch the prepared squares and aims to out-manoeuvre the corner cannon’s coverage; or the direct central soldier storm, sharpest and riskiest, covered from red’s side in the Rushed Centre Soldier guide. What red must not do is play the standard anti-screen attack by rote - the fangongma was built precisely to make that attack misfire.

The move-ten diagram rewards a careful look at black’s cannon: apparently doing nothing, actually the reason three natural red continuations are absent from the position. Prophylaxis photographs poorly; play through the next few moves and its work becomes visible.

How the system is misplayed

Treating it as a Screen Horses with an odd cannon. The two defences share a skeleton but not a nervous system: the fangongma’s back rank is thinner, its queenside slower, and a player who launches the screen’s standard counterattack timing from the fangongma’s structure discovers those differences the hard way.

Leaving the corner cannon on guard duty forever. Once red’s attack has visibly renounced the squares it watches, the cannon is just a badly-placed piece; strong players redeploy it the moment its preventive work is done, and average players lose endgames a piece’s-worth of activity behind because they never did.

From red’s chair: forcing the standard attack anyway. Pushing the river horse into covered squares, playing the memorised trading sequence into the prepared refutation - the archive is full of red players discovering that against the fangongma, the book they memorised was the wrong book.

Making it yours

Run the Counter-Palace Horse lesson on Xiangqi Arena until the move order is automatic - order matters more here than in any other black system. Then play it in slow games only, at first: the defence’s decisions are quiet ones, and blitz will teach you its shape but not its judgement.

Pair your study: for every fangongma game you review in the archive, review one Screen Horses game from the same era. The differences you notice - where each defence spends its tempi, when each counterattacks - are the actual content of both systems.