Each story walks through a single moment when the engine fired or stayed quiet — what the user saw, which factors were doing the work, and what happened next. Long-form storytelling is the antidote to opaque jargon: if the engine's behaviour can't be told as a story, it probably isn't understandable enough to act on.
Anatomy of a Framler signal — what the score is actually built from.
A walkthrough of one ticker: factor breakdown, conformal interval, regime context and calibration honesty — and what the engine does and does not claim. For live NVDA numbers, click through to /stocks/NVDA.
Read the story →Phase 3 readout that the model rated favourably — and then missed.
The honest version of a model miss: which signals fired, which didn't, what the Bayesian P(success) was, and what a confidence-weighted position would have lost.
Coming soonWhen the change-point posterior spiked before market sentiment turned.
A walk-through of one BOCPD changepoint moment — how the regime posterior shifted, which factors re-weighted, and what the dashboard surfaced over the following days.
Coming soon