Proverbs · 格言
Shogi has a stock of sayings, kakugen, that every Japanese player learns young and quotes for life: short, a little comic, and accurate. They are the game’s strategy compressed into a line each. Here are twenty-three, grouped by what they are about, with the positions that show what they mean.
The king Avoid the sitting king · An early escape of the king is worth eight moves · Close in on the king as if wrapping it · A check is only a chasing move · Against a king on the edge, the edge pawn · Never keep king and rook close
Pawns One pawn is worth a thousand gold · A pawn under the gold is harder than rock · With three pawns in hand: the joining pawn and the hanging pawn · Shogi without pawns is a losing shogi · A tokin is a gold, and more than a gold
The other pieces Drop the lance from the back rank · A knight that jumps too high is food for a pawn · Use the silver like a plover · Drop the silver at the king’s belly · The dragon in the enemy camp, the horse in your own · Drop the major pieces at a distance
Attack and defence Attack with rook, bishop, silver and knight · Defend with three generals · In the endgame, speed over material · Two for one is a trade to make, even with pawns · Drop where the opponent wants to drop · Mate with the ordinary move
Half of all proverbs are about where the king is, because half of all lost games are lost there.
A king still on its starting square is a king in the middle of the board with the rook on one side and nothing on the other. Every dropped piece near it is a threat. Move it before you attack, even one step; three steps and a silver make a castle.
When you can see the attack coming, moving the king out of its path before the first check is worth more than any defensive drop you will make afterwards. The number is a figure of speech; the point is that a king that steps away early makes eight of the attacker’s moves useless.
Do not chase with checks. Take away the squares the king would run to, from the far side first, then attack. A king with nowhere to go is mated by one gold; a king chased across the board is never mated at all.
The companion to the one above. A check that does not lead to mate usually helps the king: it runs to a better square and you have spent a move. Ask what the king does after your check; if the answer is “escapes”, find a quiet move instead.
A king in the corner or on the edge file has one way out: along the edge. Pushing your edge pawn, or dropping one, takes that square and often forces the lance or knight next to the king to move. The attack on an Anaguma starts here.
A rook beside the king is a target and a shield at once: forks of king and rook (the check-and-rook, 王手飛車) are the commonest way to lose a rook. The ranging rook openings exist partly to put the two on opposite sides of the board.
You have nine, they are worth almost nothing, and the game is decided by how you use them.
A pawn in hand attacks, blocks, lures and sacrifices in ways no other piece can, because it is cheap. Never be without one in the middlegame; trade a pawn early on the rook’s file just to have it in hand.
When an enemy rook or dragon is on your back rank, a pawn dropped on the square directly behind your gold stops it cold: the pawn is protected by the gold, the rook cannot pass, and it cannot take the pawn without losing itself. The cheapest defence in the game.
Two pawn techniques in one line. Push and exchange on a file, drop a second pawn to pull the defender’s pawn forward (the joining pawn), then drop a third one square from promotion (the hanging pawn) so that it becomes a tokin next move. Three pawns in hand are an attack by themselves.
The same idea from the other side. If your hand is empty of pawns you cannot tap, block or hang, and every attack has to be made with pieces that cost something. Look at the hands before you trade.
It moves exactly like a gold, and when it is captured the opponent gets only a pawn. So a tokin can be given away in an attack where a gold could not. The slowest attack in shogi, the “tokin crawl” up an open file, is also one of the hardest to stop.
Each piece has a proverb about how it likes to be used, and most of them are about restraint.
A lance controls everything in front of it, so drop it as far back as the position allows: it then covers the whole file rather than half of it, and it is out of reach of the pieces it attacks. The same applies to a lance defending: on your first rank, behind the king, it is a wall.
A knight cannot retreat, so a knight that jumps into the enemy camp with nothing behind it is simply attacked by a pawn and lost. Its partner proverb is 桂は控えて打て, drop the knight held back: one rank short of where it wants to be, ready to jump when the time comes.
A plover runs in zigzags, and so should a silver: forward on one diagonal, then the other. It cannot step sideways, so a silver that goes straight ahead is soon stuck; one that zigzags keeps the option of coming back.
A silver beside the king covers the squares diagonally in front of and behind it, which are exactly the squares the king would use to slip out. Protected from behind, it is the start of most mates against a castled king.
A dragon attacks best from inside the enemy position, cutting off the back ranks. A horse, with its short steps in every direction, is the best defender in the game when it stands next to your own king. Bring the promoted bishop home.
A rook or bishop from hand should land far from the piece it attacks, where it cannot be driven off by a pawn. The bishop dropped on the far side of the board that still looks at the king’s castle, the “far-seeing bishop” (遠見の角), is the classic case.
How much to spend, and where.
The attacking pieces are the two that slide and the two that leap forward without a way back. Golds stay home. An attack that uses three of these four on one part of the board is usually enough; an attack with only the rook is not.
Two golds and a silver, or a gold and two silvers, around the king. That is what every castle on the castles page is: three generals arranged so that each protects another. A fourth general in the castle is one too few in the attack.
Once both kings are under attack, the question is no longer who has more but who mates first. Give a rook to gain a move if the move mates. Counting material in the endgame is how careful players lose won games.
Trading one piece for two, a rook for a silver and a knight say, is nearly always right, because two pieces in hand are two drops. The saying exaggerates to make the point: the count of pieces matters more than their weight.
If a square would be a good drop for them, it is a good drop for you: it takes the square away and usually does the same job in reverse. Before choosing a drop, ask what they would drop if it were their move.
When closing in on the king, the plain move, the gold on the head, the pawn on the head, the obvious capture, is nearly always the right one. Brilliancies belong in tsume problems; in games the ordinary move wins because it leaves nothing to chance.
A proverb is a default, not a rule. Each one is right more often than not, which is exactly what you need when you cannot calculate: play the proverb move, and spend the thinking time on the positions where two proverbs disagree. That, in fact, is most of what getting stronger is.
The app’s review points at the move where a game turned and says why in plain words. Often the why is one of these.