Nonograms
One clue set, 120 different pictures
Investigate whether a plausible picture is actually determined by its clues. A complete solution can be valid without being the only solution.
DECISION LAB
Decision Lab: test the reason, not just the move
Nine original board investigations examine what ordinary rules leave out: missing information, shared partners, global constraints, optimality, move order and the exact goal. Make a prediction, follow the evidence, then change one condition and test the conclusion again.
Put a tempting idea to the test
Choose an investigation, commit to a prediction, then follow the evidence. Each case asks a different question: what can be proved, what remains uncertain, and which change would alter the answer. These are new reasoning cases, separate from the fixed practice studies in the learning centres.
Tents & Trees
Every tree has a neighbour. Two trees still cannot share it.
Audit a completed-looking camp using distinct tree–tent partners, then repair it without changing any clue.
Bridges
A connected network can still leave three answers
Enumerate a six-island ladder to separate three different tests: correct counts, one network and a unique solution.
Minefield
The safest square can be far from the clue
Use the board’s total mine count to prove distant squares safe without pretending to know which nearby square is mined.
Cross Lights
A working press plan is not yet a shortest plan
Exhaust all 32 first-row choices to compare four different solutions to the same diagonal-light board.
Sliding Tiles
Two moves improve the score. Only one stays on the shortest route.
Compare exact distances with a useful but incomplete tile-distance score.
Peg Solitaire
One peg anywhere is possible. One peg in the centre is not.
Separate the actual winning condition from a familiar extra condition, then enumerate every reachable finishing hole.
Four in a Row
Blocking the next threat builds the one above it
Investigate a forced loss that has no immediate winning move on the starting board and no two-column fork.
Reversi
The corner loses. The quiet move wins.
Read a complete two-empty-square ending in which move order outweighs both the corner rule and the immediate disc count.
How to read the evidence
Each case is a fixed construction or a legally reached game position. Claims about all possibilities are checked by explicit enumeration or a complete finite proof, with the scope stated alongside the result. These are authored learning investigations, not calibrated difficulty ratings or records of human playtesting. Deliberately ambiguous or invalid boards are labelled as such. Related practice links use different positions unless explicitly stated.
New to these games? The learning centres introduce the rules and individual techniques. A case may deliberately change a usual assumption or present an ambiguous position; read its setup before treating it as a standard puzzle.