Gomoku Defense vs Offense: When to Switch

Attack and Defence Are One Decision, Not Two Personalities

Ask a dozen gomoku players what separates a defensive player from an attacking one and most will describe a temperament: cautious versus bold. That framing loses games. In gomoku the first player to line up five stones wins outright — there is no score to protect and no territory to trade. Every stone you place either creates a threat or answers one, and the player who keeps creating them usually wins. The real skill is not “being aggressive” or “being solid”; it is reading the position and switching between the two at the right moment.

This guide gives you a three-question test for that switch, the counting rule behind it, and the mistakes that turn a winning position into a loss.

Why “Should I Attack or Defend?” Is the Wrong Question

The question assumes the two are opposites. They are not. A block that also extends your own line is an attacking move wearing a defender’s coat. A “safe” stone in the wrong square is simply a wasted tempo.

What actually changes hands is tempo — the right to force the other player to react. A move is forcing when your opponent must answer it immediately or lose. An open three forces. An open four forces even harder. A closed three usually does not: your opponent can happily ignore it and build their own threat while you spend another stone finishing it off.

So the practical question is never “attack or defend?”. It is: whose threats are forcing right now, and how many moves does each side still need to finish the job?

The Three-Question Test

Before you commit a stone to either job, answer these in order:

1. Does my opponent have a threat I cannot outrun? An open three qualifies. An open four means you have already lost — it should never have been allowed on the board. If a forcing threat exists, defence is no longer optional.
2. Can I do both with one move? The best square in gomoku is the one that answers their threat and extends one of yours at the same time. If such a square exists, you never had to choose.
3. If I defend now, do I still have a way back? A defensive move that leaves every one of your lines dead is a slow resignation. Prefer a block that keeps at least one of your lines alive toward the centre.

If the answers are “no threat”, “nothing that helps me” and “plenty of time”, then attack. That is the whole framework.

The Counting Rule: Which Threat Finishes First

When both sides have threats on the board, do not compare their size. Compare move counts.

• If your threat needs one move to win and theirs needs two, ignore theirs and win. This is the single most common missed win at club level: a player sees the opponent’s open three and blocks it, when their own forcing four was one move from victory.
• If both need one move, you go first only if it is your turn and that one move actually completes the line. Otherwise you must defend.
• If your threat needs two moves and theirs needs one, you have misread the position — you are not attacking, you are one move late. Defend now and rebuild.

Count your own moves, not the stones already on the board. This discipline is what separates players who win on the attack from players who merely enjoy attacking.

When Defence Is Genuinely Correct

There are three situations where blocking is clearly the right call, and none of them involve fear:

• Your opponent has an open three and you have no forcing move. Your counter-threat would need two moves; theirs needs one. Block.
• You are a full tempo behind. Their threats are arriving faster than yours. Blocking is how you buy the single move you need to catch up.
• The endgame has narrowed to a forced sequence. Once the board is tight, one loose stone is fatal; there is no room left for speculative development. Defence becomes precise rather than patient.

Notice what is missing from that list: “I feel safer blocking.” Comfort is not a strategy.

When to Switch to Attack

The signal to attack is not that you feel strong. It is that your move creates a threat your opponent cannot handle while pursuing their own plan:

• Your threat is one move quicker than theirs. Covered by the counting rule above.
• You can create two threats at once. A double threat — the fork — usually wins outright, because a single reply cannot cover both squares. Building one deliberately is the fork strategy in a sentence: never settle for one threat when the same stone or the next one can make two.
• Their “threat” is a dead three. A three with both ends blocked is decoration. Answering it is pure tempo given away. Play your own move instead and let them waste a turn trying to revive it.
• Your defensive move is also a developing move. This is the smoothest switch of all: you block, and the blocking stone becomes the first stone of your own attacking line.

Defence That Attacks: the Counter-Threat

Advanced play blurs the two categories entirely. Instead of answering a threat, you create a bigger one somewhere else, forcing your opponent to stop and deal with you. The logic is the counting rule again: if his plan takes two moves and mine takes one, I do not need to answer his at all.

This is why strong players seem to ignore obvious threats and still win. They have counted, and they are not ignoring anything — they have simply found a faster route. Learning to see those counter-threats consistently is what advanced gomoku strategy is really about; the switch between attack and defence stops being a decision and becomes a single reading of the board.

Switching Against the Computer

Against an engine the switch has a different flavour, because the engine evaluates threats more consistently than a human does. Two adjustments matter:

• It will almost always answer your forcing moves. That predictability is a tool: a forcing four does not just threaten, it steers the engine to a square of your choosing. Use forcing moves to drag it away from the area where you actually intend to win.
• It does not get bored or greedy. A human opponent may ignore your slow build; a well-made bot rarely does. Against the machine, attack only with threats that are genuinely forcing — a speculative two that no human would bother blocking may simply be ignored while the engine finishes its own sequence.

If you want to train the switch specifically, beating the gomoku AI walks through how the different difficulty levels evaluate threats, which tells you exactly which moves it will bother to answer.

Four Mistakes That Cost Games

1. Blocking the loudest threat instead of the fastest one. A big sprawling three is more visible than a quiet four-in-waiting. Count, do not stare.
2. Switching to attack one move too late. The attacking sequence was available last turn; now it is a defensive emergency.
3. Switching back to defence too early. Abandoning a live attack to block a harmless two hands the initiative straight back.
4. Defending without developing. A pure block that helps no line of yours is a tempo donation. Every defensive stone should still point somewhere.

A Practice Routine That Trains the Switch

You cannot learn this by reading. Play a game — against the AI or a friend on the board — and annotate every one of your moves with a single letter: A for a move that created a threat, D for one that answered a threat. Then review:

• Any D you played while your own threat was one move from winning? That was a lost win.
• Any A you played while your opponent had an open three? That was a lost game.
• Any move that earns neither letter? That was a wasted stone.

Do this for five games and the pattern in your own play becomes obvious — you will find you are not really a defensive or an attacking player at all, just someone who switches at the wrong moments. Fixing that is worth more than any new opening you could memorise.