|
P(blue) — chance of being empty
Fraction of consistent boards where this cell is blue.
High P(blue) = safe to click in Phase 1.
|
|
|
EV — expected SP from this click
P(ship) × E[ship value] averaged across consistent boards.
Most useful in Phase 2.
|
|
|
Score — composite scorer value
Composite score used to rank cells. Only available when source is heuristic.
|
|
|
Tooltip — per-term score breakdown
Hover over an unrevealed cell to see all scoring terms. Only available when source is heuristic.
|
The board is a 5×5 grid of 25 hidden cells. Each cell belongs to exactly one ship (or is blue/empty). Ships are horizontal or vertical contiguous segments with no overlap:
Click budget: ship hits are free. Only hitting a blue (empty) cell costs from your budget of 4 blue clicks.
Extra Chance: if you’ve hit fewer than 5 ship cells total, each blue click extends the game (the game doesn’t end). Once you’ve hit 5 ship cells, Extra Chance shuts off — the next blue you click ends the game immediately. Use your remaining budget wisely.
Perfect game: use constraint inference to find all blue cells before hitting 5 ships, then collect all remaining ship cells for free.
The solver tracks which board configurations are consistent with all your reveals so far. The Boards consistent count updates after each reveal.
Phase 1 — Avoid ships (ships hit < 5): the solver uses a weighted heuristic combining P(blue) with an information-entropy term to pick the cell that best avoids ships while reducing uncertainty.
Phase 2 — Collect SP (ships hit ≥ 5): Extra Chance is off. The solver picks the cell with the highest P(ship) × E[value]: the probability the cell is part of a ship, weighted by the expected value of that ship type.
Orange is a fixed named ship: orange cells always belong to the same dedicated 2-cell ship. Variable rare color grouping: variable rare ship cells distinguished only by forming groups are tracked by the server.