Seed recovery

Sweeps every whole second in seed_mean ± Nσ, seeds CPython's PRNG with the Oslo-local timestamp string, and weights each candidate by the Gaussian prior on the timestamp and the likelihood of your per-leg deviation observations. Runs entirely in your browser.

Parameters
Observations

One row per flown leg, in order. Each row consumes one deviation draw. Give your best estimate of the added deviation (degrees) and its certainty σ, or mark a leg blank to consume the draw without constraining it.

#deviation (°)σ (°)blank
no observations — every candidate scored on the timestamp prior alone
marginal likelihood Z =
Candidates