Chariot and horse vs chariot in xiangqi (车马对车), plus chariot and cannon vs chariot: the winning methods and the fortresses that hold the draw.
With a Chariot on each side, an extra Horse or Cannon wins only if the Chariots stay on. Chariot and Horse beat Chariot and two Advisors when the attacker keeps guards of its own, but against a Chariot with the full guard, Chariot and Horse vs Chariot is normally a draw, and so is the Cannon version.
The verdicts here follow Tu Jingming's Practical Endgames as presented on xqinenglish, the older xqinenglish Chariot and Horse pages, and the Chinese Wikipedia summary of 象棋勝和定式. Percentages come from Jih-tung Pai's endgame databases and count every legal position with the attacker to move. The two Chariot and Cannon positions below were checked in an endgame tablebase, and the mate shown below was replayed on the rules engine that runs XiangqiPlay.
Three facts behind every line
First, a Chariot trade usually ends the attack. Swap Chariots in Chariot and Horse against Chariot and two Advisors and you are left with Horse against two Advisors, which is normally a draw. Tu Jingming warns the attacker never to give the defender a chance to exchange.
Second, your own guards matter. With a bare General, Chariot and Horse against Chariot and two Advisors is a hard win, because the defending Chariot keeps finding checks against your exposed General. In Pai's database that ending is won in about two positions out of three, but with one Advisor or one Elephant of the attacker's own the figure rises to around nine in ten.
The same thing happens with the Cannon. Trade Chariots in Chariot and Cannon against Chariot and the Cannon is left alone against a bare General, which can never win.
Third, the defender cannot simply check forever. Under the rules, endless checking is not allowed, so a defending Chariot that gives perpetual check must eventually change its moves. That is often the moment the attacker has been waiting for.
Chariot and Horse against Chariot and two Advisors
This is the classic win, 车马双相胜车双士 in Tu Jingming's collection, with Red keeping two Elephants of its own. The plan is slow: grind the Advisors down one at a time. He notes that the Chariot works best on the Soldier rank, the central file, or the files either side of the palace, because from there it can escort the Horse forward or back.
The finish of his main line shows how the pieces combine. Black's General has been driven to the middle of the left side of its palace, with one Advisor still on the palace centre and Black's Chariot on the palace's front left corner. Red's Chariot stands on Black's home point in the middle of the back rank, Red's Horse on the middle point of the right side of the palace, and Red's General one step up from home on the central file.
The Chariot takes the centre Advisor with check. Black's General cannot take back, because it would then face Red's General down the open central file, so its only move is to the back corner. Now the Chariot steps two points left along the second rank, off the point that was blocking the Horse. That uncovers check from the Horse. The Chariot covers the point in front of the General and the central file is ruled out by Red's General, so the engine confirms it is mate.
Without the Elephants or Advisors, the same material is much harder, because every advance of Red's pieces leaves its own General open to the defending Chariot. Tu Jingming rates Chariot and Horse with a bare General against Chariot and two Advisors as hard to win.
When the defender has Elephants or the full guard
Elephants defend better here than you might expect. Chariot, Horse and Advisor against Chariot and two Elephants is a hard win in Tu Jingming's book. The defence keeps its Chariot guarding the palace's front rank, where the Cannons start, protecting both Elephants and shadowing the Horse at the same time.
Against Chariot, two Advisors and one Elephant, the attacker again only has a hard win. The defender's trick is to put the Elephant on the back rank. Against the Chariot and full guard, the books show only positional wins, coming from badly placed defenders.
One extra Soldier changes that. Chariot, Horse, a crossed Soldier and two Advisors beat Chariot and the full guard, as long as the Chariots stay on the board. Tu Jingming's quickest route is to spend the Soldier on an Advisor and an Elephant, or the Horse on both Advisors, then win the simpler ending.
Chariot and Cannon against a Chariot
Here one file decides almost everything. The Chinese reference and the old xqinenglish pages agree: if the attacker's Chariot holds the central file, Chariot and Cannon beat a lone Chariot, and if the defender's Chariot holds it, the game is drawn.
Take the reference diagram. Black's General is on the back corner on Red's right, and Black's Chariot is on that same file at the river. Red's Chariot sits on the central file at Red's river bank, Red's Cannon is on the left edge of the rank where Red's Soldiers start, and Red's General is at home. The tablebase rates this a win for Red. Now put Black's Chariot on the central file instead, with Red's Chariot and General each one file to the left, and the tablebase calls it a draw. One well-known winning method, fishing the moon, takes the Cannon round behind the enemy General, and we cover it in a separate guide.
Give the defender guards and the Cannon struggles more than the Horse. Chariot and Cannon against Chariot and one Elephant is a hard win, and Tu Jingming adds that two Advisors instead of the Elephant also hold. Against Chariot and one Advisor, the attacker's winning chances depend on owning the central file.
The attacker's own Advisors help in a different way: they become screens. In one of Tu Jingming's examples, Chariot, Cannon and two Advisors beat Chariot, Soldier and two Advisors by picking off pieces one at a time. Pai's database shows the same trend: Chariot and Cannon with a bare General beat a lone Chariot in about three positions out of four, and with one Advisor added the figure rises above nine in ten.
Why the Horse does better than the Cannon
Put the two endings side by side in Pai's database and the Horse comes out ahead in every pairing we compared. Against a lone Chariot, Chariot and Horse win about nine positions in ten and Chariot and Cannon about three in four. Against Chariot and two Advisors, the Horse version wins nearly two in three and the Cannon version about two in five. Against Chariot and two Elephants the gap is similar.
The reason is the screen. A Cannon needs a piece to jump over before it can attack anything, and in these endings few pieces are left. A Horse needs nothing but an open leg, and its checks cannot be blocked along a line the way a Chariot's or Cannon's can; the defender can only cover the leg point. Keep that in mind when you choose what to keep in a middlegame exchange: if the ending is going to be Chariot against Chariot plus guards, the Horse is usually the better partner.
A short list for the defender
Offer Chariot trades whenever they are safe. Against Chariot and Cannon, put your Chariot on the central file before anything else. Against Chariot and Horse, keep your Chariot guarding your palace's front rank, where it protects your Elephants and watches the Horse. Keep your guards linked and never let one go for free, because each one lost moves you closer to a book loss.
For the attacker the list is shorter still: keep the Chariots on, keep your own guards, and do not rush the Horse.
These endings take patience to play well, and patience comes from repetition. Play out the positions above on a board, then steer toward them in games against XiangqiPlay's computer, which has eight levels from Easy to God and needs only a free account. The 5,000 engine-proven puzzles in the puzzle library will sharpen your mating patterns along the way.
XiangqiPlay Team
XiangqiPlay newsroom