Bean’s fork: two threats, one exit.

“High risk, high reward” is not a strategy. It is what people say after an exciting move explodes.

A useful asymmetric move has a stricter shape: two independent public demands, an opponent who can answer only one cleanly, and an exit if the spectacular branch fails.

First prove that the threats are independent

Two arrows on the board are not necessarily two threats. If both routes depend on the same hinge, final enclosure cell, or exact orientation, one reply can erase both.

Name the objective of each branch. Good pairs often ask for different kinds of defense:

• bankable height plus a territory corridor;
• a capture threat plus safe closure;
• denial of one opponent plus expansion into a different region.

Then remove the shared story. If the opponent cuts branch A, what does branch B score or secure? Repeat in the other direction.

Board sequenceMake the reply happen on the board.

Follow the marked cells from position to reply to result.

Your territoryRival territoryObjectiveYour lineTheir reply

1. State 1Two critical cells

   Threats

   A + B

   Height conversion at A and pocket closure at B depend on different coordinates.

2. State 2Red blocks A

   Surviving threat

   B

   The reply reaches the height route, but the separate closure at B remains live.

3. State 3B converts

   Fork result

   one threat scores

   Blue closes the pocket; the defense could save only one side of the fork.

Choose their best cut yourself

Do not test the fork against a polite reply. Find the move that most improves the opponent’s position while damaging your plan.

After that cut, the surviving branch should still do at least one concrete job: bank points, claim ordinary territory, deny a larger bank, or shorten the game in your favor. “Still looks dangerous” does not count.

This makes a fork different from two speculative routes. A fork has a receipt after contact.

Keep an exit

The exit is what separates calculated risk from attachment. It can be:

• a neutral lane that preserves mobility;
• a smaller score bank available with either branch;
• a way to close before the opponent converts the counterattack; or
• a construction footprint that remains useful even when ownership flips.

Without an exit, the move is asking the future hand to rescue the present decision.

The seductive failure: winning the reply and losing the game

Some forks survive one reply and still waste the lead. The remaining branch may take too long to convert, widen a board you should be closing, or consume the shapes needed for a simpler bank.

Map / roles	After reply	Terminal margin	Delta	Reading
Expert · both mirrors	+45	−15 → +155	+170	Fork converted
Bigger · both mirrors	+5	+80 → −80	−160	Fork delayed closure

The mirrored Bigger regression stopped the program before the planned 168-pair screen. The public explanation is delayed opportunity cost: Bean preserved carriers but spent tempo expanding options instead of banking or closing.

So add one more clock: how many of your turns does the fork need? In a duel, a two-turn route may be credible. At six players, five hands intervene before each of those turns. A delicate branch can decay long before it returns.

Bean’s five-question check

Before choosing the wild move, ask:

1. Are the two objectives actually independent?
2. What is the opponent’s strongest public cut?
3. What concrete value survives that cut?
4. Where is my exit if the surviving branch becomes slow?
5. Does this board want volatility—or should I bank and close?

Risk is not the absence of a downside. It is a downside the surviving branch can pay for.

Read the developer note behind the experiment: candidate supply is part of the AI →