Oracle
The prediction layer. Oracle turns the statistical signals already visible in the Terminal into forecasts that carry a published accuracy record from the first day they appear.
Why it comes last
Prediction was deliberately not the first thing built. A forecast is only as good as the observation beneath it, and an unverifiable forecast is worse than none — it converts a guess into something that looks like knowledge, which is harder to argue with and easier to act on.
So the order was: collect and attribute the observations, make the dependencies auditable, build a harness that can tell whether a signal actually preceded anything, and only then start forecasting. The harness exists and runs nightly. It is what produces the accuracy record already visible in the Terminal.
What it will do
- Specific, checkable outcomes. A corridor's traffic falling below a threshold inside a stated window — not a vague sense of elevated risk.
- Recorded before the fact. Every forecast is written down before the window opens, scored after it closes, and never quietly revised afterwards.
- Willing to say nothing. Where the evidence is too thin, it declines to forecast rather than producing a confident-looking number nobody can check.
What it inherits
Oracle does not start from scratch. The signal miner, the point-in-time replay, the ground-truth ledger and the scoring harness are all live today, and every one of them was built so that a forecast could later be checked against them without rewriting history.
The bar it has to clear
A prediction that cannot be scored is marketing. Oracle ships when its forecasts beat a null baseline on a sample large enough to say so — and it will publish the sample size and the misses alongside the hits, exactly as the Terminal already does.
Until then, everything in the Terminal marked as model output is labelled as such, and nothing there is presented as a forecast.