OverGrid punishes a familiar board-game habit: scoring your turn at the moment your hand leaves the piece.

A construction is not worth its immediate total. It is worth what it banks now, what it leaves open, and what the next player can do with the public board you created. You do not need to predict their unrevealed hand. You need to test the strongest reply the visible position already promises.

Use reply-testing only when it can change the choice

If one construction clearly scores more, closes safely, and exposes less, play it. Calculation is most useful when two candidates are close.

For each close move, ask:

  1. What is the opponent’s best public objective: capture, height, closure, or denial?
  2. Which visible hinge, corridor, or loose region makes that objective possible?
  3. Can their reply erase what I thought I banked?
  4. After that reply, which of my routes still exists?

The point is not to imagine the cruelest move in the universe. A rational reply should improve their position, not merely hurt yours.

Board sequenceCompare the board after the reply, not before it.

Follow the marked cells from position to reply to result.

Your territoryRival territoryObjectiveYour lineTheir reply
  1. State 1Create the threat
    Create the threatBlue prepares a capture chain that will harden the marked pocket next turn.
    If ignored
    +30 next turn

    Blue prepares a capture chain that will harden the marked pocket next turn.

  2. State 2Reply is forced
    Reply is forcedRed abandons its preferred height route to occupy the only closing cell.
    Red’s preferred route
    delayed

    Red abandons its preferred height route to occupy the only closing cell.

  3. State 3Spend the tempo
    Spend the tempoBlue uses the stolen turn to advance a separate objective while red remains tied to defense.
    Initiative
    blue keeps it

    Blue uses the stolen turn to advance a separate objective while red remains tied to defense.

Read this asBlue's threat is valuable because Red must abandon a preferred height route. The payoff is the separate objective Blue reaches while Red is occupied.

Separate loss from disappointment

Suppose Move A earns 30 now and gives the opponent a clean 25-point capture. Move B earns 20 and leaves only a 5-point reply. Move A is still not automatically wrong: its footprint may create a second bank next turn. But its honest comparison starts at a net swing of 5, not at the headline 30.

That tiny ledger prevents three common mistakes:

  • counting territory you have not enclosed yet;
  • treating printed height you merely approached as scored;
  • ignoring that a capture changes both players’ totals.

In a duel, test one reply. In a crowd, test resilience.

Heads-up play rewards precision because the board comes directly back after one opponent acts. At a four- or six-player table, several turns will rewrite the frontier. You usually cannot calculate an exact sequence, so ask a different question: does the construction remain useful if any one of its visible promises is interrupted?

A resilient multiplayer move has more than one way to convert. It may score now and preserve mobility, threaten height from two sides, or narrow neutral territory without relying on a particular player to cooperate.

The twenty-second receipt

Before committing a close decision, finish this sentence:

I prefer this construction because after their best visible reply, I still ________.

Good endings are concrete: “hold the hinge,” “bank the height,” “close two targets,” or “keep both routes.” If the blank can only be filled with “have pressure,” you have not finished calculating.

Next player guide: a route is not a bank →