Data Sources

User traffic never reaches these providers directly. A background worker fetches on a fixed schedule, the results are cached on our own infrastructure, and every page reads from that cache — which is why the site stays up even when a provider does not.

Price history (candles across 1-hour, 4-hour and daily timeframes, in the millions of rows), detected patterns with their full quality-gate verdicts, indicator events with measured outcomes, computed scores with their component breakdowns, and aggregated news. Derived statistics — success rates, baselines, correlations — are computed from this store, not fetched from anyone.

Every value carries the label of its class: near real-time, delayed, daily reference, or historical. The classification comes from the feed that produced the value, not from its category — the same coin reads differently depending on which feed answered — and the status is recomputed each time a page is served rather than stored with the record. A feed that stops updating stops being described as current: past its window the value is still shown, labelled stale and stamped with its age.

No feed on the platform streams, so nothing here is labelled “live”. The fastest is a three-minute poll. A daily ECB reference rate is never presented as a quote, and a closed futures market is shown as closed rather than as a failure. The full rules are on the methodology page.