Download the corpus
The whole graded set of rules this planner runs on is one file you can download, read and keep: 136 rules, each with the mechanism it works by, the strength of its evidence, whether a human has checked its sources, and those sources.
What is in it
One entry per rule, in machine-readable form: the claim, the named mechanism, the evidence grade, whether the sources have been verified by a person, the sources themselves and, when the file is built from a full clone of the history, the dated record of every change that rule has been through. A stable identifier on each rule lets you cite one and find it again. The file names the exact commit it was built from, so two people who download it can tell whether they are looking at the same corpus.
What the grades mean
Every rule carries one of five grades, from “well established” down to “refuted”, and the file spells each grade out in those words. A rule is graded “well established”, “promising” or “a good hunch” only if it can state its mechanism in one sentence; a claim that only correlates is downgraded or dropped. The file also holds the refuted tier, 4 folk claims this corpus checked and rejected, because a correction is worth as much as a rule.
Every rule is marked verified or unverified
127 of the 136 rules were checked against a source by a person, and 8 were settled by arithmetic, geometry or a definition. Anything a language model asserted but no person has yet checked is marked unverified and never presented as fact, and the model is never allowed to promote its own work.
The license
The corpus is offered under Creative Commons Attribution-ShareAlike 4.0. You can use it, build on it and ship it, as long as you credit it and pass on the same freedom. Share-alike is deliberate. We don’t want a mechanism-graded rules file absorbed, uncredited, into a planner that won’t admit its companion charts are folklore.
For teachers and extension educators · Read the rules on the site · What we don’t know