About Bluebird

Bluebird tests one narrow question: which forecast grids match one published seven-day weather profile?

Why it exists

This private prototype compares fifty ECMWF ensemble scenarios across a curated set of SFO nonstop destinations. Every member becomes Strong, Borderline, or Poor; only decisive Strong or Poor support produces Go or Don’t. Everything uncertain becomes Hold.

What it is not

Bluebird does not judge whether a destination is generally good, and a gateway city’s forecast does not prove conditions in nearby mountains, beaches, or parks. It does not check fares, seats, schedules, hotels, crowds, accessibility, air quality, humidity, smoke, or personal interests. The flight-reach bands are coarse planning estimates, not schedules or availability.

Inspect the method

The standalone engine explanation shows every used input, excluded input, threshold, tie-break, and fail-closed state. The accuracy ledger preserves completed-run evidence without rewriting old calls.

How to reach us

Questions, corrections, and wrong-call reports go through the monitored contact form.