Corner and Edge Strategy

The AI now considers adjacency and defense, but it treats all board positions equally. A corner cell scores the same as a centre cell if they have the same neighbours. But in territory games, corners and edges are strategically different — they have fewer adjacent cells, making them easier to secure.

This unit adds positional awareness. Corners get a +2 bonus, edges get +1. The AI now prefers strategic positions, especially in the opening when few cells have neighbours.

Run It

pasmonext --sna inkwar.asm inkwar.sna

Watch the AI’s opening moves. It now gravitates toward corners and edges rather than claiming random centre cells.

Why Corners Matter

Consider a corner cell at position (0,0):

• Only 2 adjacent cells: (0,1) and (1,0)
• Easier to “secure” — fewer cells needed to surround it

Compare to a centre cell at (3,3):

• 4 adjacent cells: up, down, left, right
• More contested — more ways for opponents to claim adjacent territory

In games like Reversi/Othello, corners are famously valuable because once claimed, they can’t be flipped. Our game doesn’t have flipping, but the strategic principle remains: fewer neighbours means less exposure.

Position Bonus Calculation

The new function determines if a cell is a corner, edge, or centre:

The logic:

1. Extract row and column from the cell index
2. Check if row is 0 or 7 (top/bottom edge)
3. Check if column is 0 or 7 (left/right edge)
4. Corner = both row AND column are edges
5. Edge = either row OR column is edge (but not both)
6. Centre = neither

Combined Scoring

The position bonus adds to the existing adjacency score:

Three factors now contribute to each cell’s value:

• Adjacent AI cells (offense)
• Adjacent human cells (defense)
• Position bonus (strategy)

Opening Game Impact

The position bonus has the biggest impact early in the game:

Without position bonus (Unit 11):

• First AI move is random (no neighbours yet)
• Any empty cell scores 0
• No preference for strategic positions

With position bonus (Unit 12):

• Corner cells score 2 even with no neighbours
• Edge cells score 1
• AI prefers corners > edges > centre

As the game progresses and cells accumulate neighbours, adjacency scores dominate. But the opening moves now show strategic awareness.

The Board Layout

Visualising the position bonuses:

2 1 1 1 1 1 1 2    <- corners (2) and top edge (1)
1 0 0 0 0 0 0 1    <- left/right edges (1), centre (0)
1 0 0 0 0 0 0 1
1 0 0 0 0 0 0 1
1 0 0 0 0 0 0 1
1 0 0 0 0 0 0 1
1 0 0 0 0 0 0 1
2 1 1 1 1 1 1 2    <- corners (2) and bottom edge (1)

24 edge cells (including 4 corners) get bonuses. The remaining 40 centre cells get none.

The Complete Code

Try This: Higher Corner Value

Make corners even more attractive:

.gpb_corner:
            ld      a, 4                ; Corner bonus = 4
            ret

.gpb_edge:
            ld      a, 2                ; Edge bonus = 2
            ret

This makes the AI aggressively pursue corners, even over contested centre cells.

Try This: Centre Preference

Or flip the strategy entirely:

get_position_bonus:
            ; ... extract row/col ...
            ; Check if centre (not edge)
            ld      a, b
            or      a
            jr      z, .gpb_edge
            cp      7
            jr      z, .gpb_edge
            ld      a, c
            or      a
            jr      z, .gpb_edge
            cp      7
            jr      z, .gpb_edge
            ; Centre - give bonus
            ld      a, 2
            ret
.gpb_edge:
            xor     a
            ret

This creates an AI that prefers the centre, leaving corners for later.

What You’ve Learnt

• Positional evaluation — Static bonuses based on board location
• Multi-factor scoring — Combining adjacency + defense + position
• Opening strategy — Early moves now show preference
• Edge detection — Checking if values equal 0 or maximum

What’s Next

In Unit 13, we’ll add difficulty levels — letting players choose between easy (random), medium (adjacent), and hard (full strategy) AI opponents.

What Changed