NIK · THE TESLAQUAKE EVALUATOR
Every claim that some factor moves the earthquake count gets the same treatment: the agency's own data, a stated null model, declustered catalogs, a fixed seed — and the result is published whether it finds something or nothing.
- 1
- Evaluations
- 1
- Null
- 0
- Effect
- 0
- Inconclusive
- 3
- Files fingerprinted
Bring us your question
A spreadsheet you’ve kept for years, a trunk of your grandfather’s records, a public file nobody has checked. One question, one dataset, and we test it the same way we test our own — pre-registered, seeded, reproducible, published either way.
We publish the method before the result, the data with the result, and the misses with the hits.
The board
One row per evaluation. NULL means the factor did nothing measurable; EFFECT means it did and the page says how much; INCONCLUSIVE means the sample was too small to say. Every row links to its method, its numbers and the command that reproduces them.
Evaluate → Certify → Publish
Evaluate
NIK, the lab's always-on machine, downloads the agency data, declusters the catalog, draws the null from a fixed seed and writes the result file with a SHA-256 for every input.
Certify
The result, the verdict and the chart are hashed and time-stamped (OpenTimestamps), so the published numbers can be proven unchanged later.
Publish
Five gates must all pass — primary sources, reproducibility, fingerprints, stated method, clean language — before this page is even built. There is no override.
Take the data, not our word
The board above is rendered from evaluations.json — the only file the site reads. Want a factor tested? Start at Submit an idea; the terms are on that page.
TeslaQuake sells a $3.69/month membership to its live tools. The Evaluator is free, and its results are published regardless of what they show.