Hey HN,
I'm a solo dev. I built this API to solve a problem I saw in both finance and engineering: Non-Deterministic Data.
Everyone measures accuracy (RMSE), but few measure Predictability (how stable is the variance?).
The engine calculates a 'Predictability Score' (0-100) using a tunable volatility constant (K-Factor).
Tech Stack: Flask, Numba (for speed), Postgres.
Use Case: Detecting sensor drift before failure, or catching AI hallucinations before they hit production. It's live at https://www.predictability-api.com. Would love feedback on the math or the API design.
Thanks. Ryan