The questions we expect, and the ones playtesting raised, answered plainly.
Sessions during development kept forcing rebalances, which is what fun pressure looks like from the inside: a player loses in a way that feels wrong, argues, and the mechanism has to answer (the alignment-earned model and the hard-crossing rebalance both started that way; the commits are in BALANCE.md). Playtesting so far is mostly bots plus a handful of humans, so treat this as a claim under test. Runs take 20 to 40 minutes, losing is common, and the debrief is built to make losing worth it. The public alpha exists to test that against people who did not build the game.
A game. It does not tell you racing is wrong; it puts you in the seat and lets the incentives do what they do. Three worldview presets (cautious, consensus, skeptic) carry genuinely different assumptions with their own sources, and the game refuses to say which one is true. If one side of the real debate thinks a preset misrepresents them, that is a bug report we want.
Because the alternative is worse. A game has numbers no paper can back: turn counts, hand sizes, thresholds. Those cite the design constitution and wear a design badge, so no game-feel number ever hides behind a fake empirical citation. They are counted separately from external sources everywhere counts appear.
No, and we say so everywhere it matters. The machine proves every number carries a citation, that citation tiers are honest both directions, and that every published count agrees with the data. Whether a source truly supports a value is human judgment; EVIDENCE.md shows worked examples of how figures became game values, and the challenge-a-number issue template exists because we expect to be wrong somewhere.
Probably because racing blind loses, which is the thesis, not a difficulty bug. The full outcome distributions are published in BALANCE.md, counter-paths included. If you think the corridor to a good ending is too narrow to be fun, that is exactly the feedback the alpha is for.
By a very small human core with heavy AI assistance, quickly, and the git history
shows it. The complete account is in GOVERNANCE.md under "How
this project is built." The trust architecture exists precisely so that nothing
rests on trusting the authors: run pnpm validate, read the map, file a challenge.
We mostly do not, and we accept that. No analytics ever ships. What we count is what people give freely: debrief feedback, shared seeds, issues, emails, and the hosting provider's ordinary aggregate request counts, which involve no code on your device.
That banner is what GitHub shows every logged-out visitor on every repository (compare any large project while signed out). Signed in, the templates are open. No account? Email criticalwindow@chipmates.ai; the debrief links it too.
Planned, not shipped. Its groundwork already accrues during play, and it ships after the alpha, once it can land unspoiled.
Because someday someone will host a modified fork, maybe with a backend bolted on. AGPL means whoever plays that fork over a network is entitled to its source too. The license protects the players of every future version, not our server, of which there is none.