Defend without conceding the game

Before looking for an attack, check whether the opponent has a valid line. If not, you may choose any opposing tile to remove and assess which structure to dismantle. If they do, the situation changes: you can continue only if your removal breaks every completed line. A developing threat does not have the same priority as a finished line.

Imagine a red straight on A1–D1 and a promising red trio on F4–H4. Removing from the trio might look like a good defensive investment, but it leaves the straight intact and gives Red an immediate win. Your candidates this turn must come from A1, B1, C1 or D1, provided no other red line makes the choice more specific.

If there is exactly one opposing line and no other, any of its four tiles breaks it. No rule requires removing from the center. With A1=2, B1=3, C1=4 and D1=5, removing an endpoint or an interior tile eliminates that window. Choosing among these legal defenses is the second part of the problem.

Among removals that save the game, consider the continuations: which cell opens, which other opposing projects disappear and which number returns to their reserve. These are strategic criteria, not a mathematical guarantee of the best move. The proven requirement is to break the line; the long-term comparison depends on the whole position.

With several lines, breaking only the most obvious one is insufficient. Look for a tile that belongs to all of them. If a red horizontal and vertical line cross only at D4, removing D4 destroys both. Removing another horizontal tile leaves the vertical intact and loses immediately, even though the first line disappears.

You can perform this operation without formulas: list the first line’s cells, cross out those absent from the second and repeat for the remaining lines. Cells surviving every check are your valid defenses. If none remain, there is no saving removal; the opponent has built a position that survives every choice.

You prove a defense by looking at what remains after removal. Scan rows, columns and diagonals to make sure no other valid window survives. This habit matters particularly with straights of five or more tiles: breaking one window may leave another window of the same long run intact. The number of lines destroyed is not enough.

You cannot solve the defense by placing a tile afterward in an empty cell of the opposing line. The victory check would already have happened. While calculating, mentally separate removal from placements. First require a board with no opposing lines; only then study the two tiles that will start your own plan.

In defensive problems, try to justify why the alternatives fail. “I remove D4 because both lines share it” is a checkable answer; “it looks like the most dangerous tile” does not identify the deciding condition. When several answers are correct, do not invent a single mandatory cell: recognize every removal that eliminates the lines.

Then distinguish obligation from preference. The obligation is to prevent immediate defeat: find a cell belonging to every opposing line. The preference is to leave yourself better prospects when several defenses meet that obligation. Then compare which structure you dismantle, which cell opens and which number the opponent recovers, without confusing a valid defense with a necessarily optimal move.