The Single Horse Defence

One horse up, one horse wide

The edge horse looks like the concession and is actually the plan. By declining the second central post, black keeps the queenside file clear for the chariot, gives the rerouted cannon a clean lane to the advisor-corner file, and removes the target red’s standard levers aim at - there is no second screen horse to undermine. The centre soldier’s defence passes to the central horse and the elephants, a burden the shape is specifically built to carry.

The system’s temperament follows from its geometry: black’s setup is complete early, hard to provoke, and short on the tactical handles red’s main attacking schemes grip. Red usually obtains the freer game - the model line shows red’s cannon infiltrating the very file the edge horse left thin - and black’s wager is that free-looking and profitable are different things against a position with no targets.

Play it if the Screen Horses feels crowded to you: the dantima is the defence for players who would rather hold four strongpoints than eight squares of theory.

The grand old main line

The single horse setup is one of the classical era’s principal defences - the old manuals treat it as a peer of the screen formations, and for the first decades of the twentieth century it remained standard equipment at the highest level, associated especially with the southern schools’ patient stylists. Its theory is old enough to have furniture: named sub-variations, stock traps, whole endgame tabias.

Mid-century analysis judged it a shade passive against the newly-systematised cannon batteries and it drifted from main-line status - which is precisely what makes it valuable now. In the archive’s era it appears as the educated surprise: a defence with a hundred years of worked-out answers that most modern opponents have never had to face across the board.

The model line, move by move

The first five steps could be any classical game; the system announces itself at step six when the queenside horse takes the edge instead of the screen square, and step eight completes the idea - the cannon rerouted to the advisor-corner file, the dantima’s signature battery. Red responds the critical way, driving a cannon deep into the file black’s horse left, and the walkthrough ends with the defence calmly reinforcing.

Notice what black never does in these twelve moves: touch the kingside structure, loosen a soldier without cause, or answer red’s infiltration with panic. The system in one sentence is 'nothing loose, nothing hurried'.

The ideas underneath

No handles: the system’s first idea is negative space. Red’s prepared attacks against classical defences work by gripping something - a screen horse to lever, a cannon to trade against structure. The dantima’s oddness removes the grips, and red’s opening advantage must be built by hand, move by move, against a position that keeps offering nothing to hold.

The advisor-corner cannon is the shape’s engine room: it guards the palace approaches, supports the centre laterally, and - the part opponents forget - becomes the counterattack’s rear battery the moment red overextends. Most dantima wins run through that cannon firing along a file red assumed was purely defensive.

The edge horse’s day comes late. It holds the wing quietly through the opening, then reroutes - via the advanced wing soldier’s road the model line opens - into exactly the river squares red’s middlegame commitments leave unguarded. The piece is not misplaced; it is scheduled differently.

Plans for both sides

Black’s plan has three chapters: complete the fortress (horses placed, corner cannon set, elephants linked - the walkthrough’s arc), absorb red’s freer development by giving its activity nothing to buy, then counter in the middlegame’s second act - the corner cannon’s file, the rerouted edge horse, and the untouched kingside majority are the standard instruments. Dantima games are won in chapter three by players who did not improvise in chapter two.

Red should take the concession honestly: more space, freer pieces, the infiltrating ideas the model line shows - and then convert by patient accumulation rather than assault, because assaults need handles and there are none. Red’s best games against this defence look like slow constriction; red’s worst look like a storm breaking on a rock, followed by the corner cannon’s counterattack.

The move-ten diagram repays study from both chairs: red’s position is visibly freer, black’s visibly sounder, and neither claim is false. The defence asks black to believe the second claim matters more - a hundred years of its games say that faith is well-placed more often than the eye suggests.

How the old machine is mishandled

Playing it half-heartedly. The dantima’s pieces defend as a set - edge horse, corner cannon, linked elephants each covering the others’ blind spots. Substituting one 'improvement' from modern setups (a screen-style horse here, a delayed cannon there) leaves holes the original never had; the antique works because all its parts are original.

Answering infiltration with material. Red’s advanced pieces tempt black into winning them at structural cost - chasing the deep cannon with soldier moves, trading a fortress piece for the intruder. The defence’s economics say let it sit: an infiltrator attacking nothing is red’s tempo problem, not black’s.

From red’s chair: declaring victory at move ten. The freer game against the dantima is the opening’s rent, not its purchase - red players who stop improving and start attacking on schedule find the fortress intact and their own commitments due. Against this defence the middlegame’s first question is always 'attack what, exactly?'

Making it yours

Learn the setup in the Single Horse Defence lesson on Xiangqi Arena, then give it a dedicated run in rated play - ten games minimum before judging it, because the defence’s rhythm (quiet, quiet, counter) takes exactly that long to trust.

Study its archive games with the corner cannon as your protagonist: note the move it stops defending and starts attacking in black’s wins. That single transition - chapter two into chapter three - is the entire art of the system.