Exhaustive search workbench

Mining order N = 8

Probe the complete row-balanced universe. Each rejected prefix can eliminate an exact block of boards; every survivor is checked with integer directions across every slope.

Possible positions Loading exact search space…
Tried directly 0 Boards explicitly inspected
Culled by proof 0 Boards skipped by invalid prefixes
Solutions found 0 Exact 2N-point witnesses

Exhaustion meter

0 positions classified

0%
No positions classified yet. Frontier depth 0 / 8 rows

Live search field

N = 8 lattice

Current score board on the grid
Ready to prospect Run a batch to watch sampled candidates cross the grid.
Current signal No candidate yet

The animation samples the worker; rendering never sits inside the hot search loop.

Cut ledger

What the engine rejects

0 batches
Repeated direction
0
Column overflow
0
Exact survivors
0
  1. Create a batch to start the local search ledger.

Strike cabinet

Found boards

0 stored

Exact solutions found while mining this N will be pinned here.

Engine anatomy

Small proofs, stacked early.

The meter is not an estimate. It advances only when a board is checked or an entire lexicographic subtree is ruled out by a necessary condition.

01

Forced balance

At 2N points, no-three-in-line forces two dots in every row and every column. Rows are encoded up front; column capacity prunes prefixes.

02

Integer directions

Normalized primitive vectors detect repeated slopes at each anchor. There is no floating-point geometry and no missed rational slope.

03

Prefix certificates

If the first k rows already contain a triple or a third point in one column, every completion of those rows is impossible and is culled together.

04

Guided repair

Two-permutation swaps preserve exact row and column balance. Guided mode aims them at conflict-heavy rows, keeps improvements, permits occasional escape moves, and restarts when trapped.

Scope: This browser miner is an exact census engine for its stated row-balanced universe, but it is intentionally lightweight. It does not reproduce the specialized repair, Hall-matching, SAT, or portfolio solvers in the N=71 and N=81 investigations.