Our larger tic-tac-toe boards both require four consecutive marks, but play differently. On 4×4, a valid line spans the board in its direction. On 5×5, additional offset four-cell windows appear. The original examples below treat both cases precisely.
In this guide
Separate board size from winning lines
Count a winning line as a specific sequence of four cells. On 4×4 there are four horizontal, four vertical and two diagonal lines, totalling ten. Each row of 5×5 has two four-cell windows, as does each column, with eight diagonal windows as well: 10 + 10 + 8 = 28.
More possible lines do not automatically imply a forced first-player win. The counts describe geometry. To make a decision, next check which lines are already blocked by opposing marks.
| Board | Lines and target |
|---|---|
| 3×3 | 8 lines · three marks |
| 4×4 | 10 lines · four marks |
| 5×5 | 28 lines · four marks |
4×4: a blocked three has no spare cell
A 4×4 row reading X–X–X–O can no longer win for X. There is no fifth cell into which the window could shift. The same applies to a full column or main diagonal. Playing along such a dead line can help only through other crossing directions.
Do not merely collect neighbouring marks. Find intersections of lines that remain available. A setup square is useful when it prepares several unblocked finishing routes without allowing an immediate opposing four.
4×4: create two distinct finishing cells
In this study, X occupies b1, a2, c2 and b3. O occupies a1, d1, d3 and a4. X can play b2. Then the second row needs only d2, while column b needs only b4. The two cells differ, so O cannot block both in one move and has no immediate win here.
The four earlier X marks are not all in one long row. Their connection appears at b2. This is the key lesson: on 4×4, prepare a second direction rather than attaching every move to the same existing pair.
Position as text
X: b1, a2, c2, b3. O: a1, d1, d3, a4. To move: X. Rows count from the top.
5×5: an open three can be decisive
Consider row three as empty–X–X–X–empty. It has two immediately available winning cells: a3 completes a3–d3, while e3 completes b3–e3. Only three X marks are present, so the game has not ended, but one ordinary block is insufficient.
Both ends must be empty, and opposing counterplay must be checked. At the edge, or with O at one end, often only one finish remains. If O already has a winning move, O can win immediately instead. An open three is a specific tactical structure, not an unconditional guarantee.
Position as text
X: b3, c3, d3. O: a1, c1. To move: O. Rows count from the top.
5×5: a gap can still be an immediate threat
Three adjacent X marks are easy to spot. X–X–empty–X inside a four-cell window is less conspicuous but also needs only one move. In this study, X at a4, b4 and d4 wins by taking c4. Playing e4 merely adds a mark without closing the gap.
Check every four-cell window in both directions. Look for exactly three own marks and one empty cell, whether or not the three already touch. Apply this horizontally, vertically and diagonally.
Position as text
X: a4, b4, d4. O: b1, d1, c2. To move: X. Rows count from the top.
5×5: a diagonal need not start in a corner
b1–c2–d3–e4 is a valid four-cell diagonal, touching neither a1 nor e5. Scanning only the two corner-to-corner diagonals misses it. In this study, X at b1, c2 and d3 needs e4; O cannot rely on a1 having blocked the long main diagonal.
Position as text
X: b1, c2, d3. O: a1, e1, a5. To move: X. Rows count from the top.
Two windows are not always two threats
Row three has X at a3, b3, d3 and e3. Both four-cell windows need c3. These are two almost-complete lines but only one distinct winning square. O can occupy c3 and stop both at once. This differs from an open three whose two endpoints are different cells.
Position as text
X: a3, b3, d3, e3. O: a1, c1, e1. To move: O. Rows count from the top.
Avoid a blanket centre rule
On 5×5, c3 belongs to eight four-cell windows and a1 to three, explaining the centre’s flexibility. On 4×4, however, b2 belongs to three winning lines and a1 also belongs to three. The geometric comparison is not identical across sizes.
During play, count only lines still available. A central cell with several opposing blocks can be less useful than an edge cell creating a direct win or genuine fork. Concrete tactics outrank a general location preference.
A routine for every large-board move
First find every own four-line finish. Next identify opposing finishes and their distinct cells. Then distinguish genuine forks from threats sharing one block square. Finally compare several still-open windows before selecting a setup move. After placing it, repeat the check from the opponent’s perspective.
Practise 4×4 and 5×5 separately with another person on one device. Name your intended winning window before moving. If a calculation fails, replay the short study sequence in a fresh round. The diagrams are explanations, not positions automatically loadable into an editor.