

The pitch is “six degrees of Kevin Bacon as a golf course”: each hole is a pair of people — say, Ryan Gosling to Harrison Ford — and every film you name is a stroke. The interesting design decision was rejecting route commitment. Early versions had you pick a path up front; the version that survived playtesting is free-shot — you just name a film, the game reveals what links it opened, and the pre-built routes exist only as a read-only course map. That turned out to matter a lot: planning a full route in advance is nearly impossible, but discovering the path shot-by-shot, being forced in a direction and finding the payoff, is exactly the kind of difficulty that feels rewarding rather than obscure.
Under the map is a real graph: IMDb’s public TSV dumps ingested into ~46k films,
~219k people, and ~600k typed edges (actors, directors, cinematographers,
composers), with a BFS that measures distance in strokes rather than hops.
Building it revealed the difficulty dial I’d been looking for: restrict
endpoints to the top 1,000 most-connected people and almost every pair collapses
to two strokes; widen the pool to 20,000 and most pairs need three, with a long
tail at four-plus. So difficulty scales with navigation distance through a
recognizable pool, not with obscurity — a lesson learned the hard way after a
playtest hole ending on a Bollywood actor most players couldn’t place. Holes are
generated from the graph, then curated; the solver’s validChains output is the
single source of truth the game client checks shots against.
Roadmap
- Nine-hole prototype on Netlify
- Scorecard, shot log, free-shot play + course map
- Daily-puzzle shell
- Shareable emoji scorecard polish