This is not a single widget—it is a fused pipeline: Weinstein-style stage labels, ML-derived forward stage distributions, SATA-aware validation, and ensemble checks before confidence is surfaced. Retail investors get plain-English stages and signals; practitioners see the same tensors, probabilities, and gates we use in production analysis and screening.
Cross-service intelligence
TRINITY threads price and liquidity from Hermes-style caches, runs Athena's stage pipeline (including ML where enabled), enriches screening via Dionysus market APIs, and lands everything in Iris with anchors for signals, smart-stage reads, and ML prediction blocks. What you explore on a symbol page is the same analytical substrate that powers filters and rankings elsewhere.
Plain-language on the surface; probabilities, gates, and shared schemas underneath—the same concepts traders screen with map cleanly onto institutional workflows.
Symbols carry stage descriptions and numeric confidence that reconcile across backfill snapshots and live reads—so stage strength stays comparable day to day.
The symbol ML prediction panel surfaces explicit distributions across forward stages—not a black-box sentence—so sophisticated users can compare model mass to price action.
to_stage_1 … to_stage_4 style probability payloads propagated through Iris APIs.Athena's pipeline couples SATA scores with stage decisions—when structure disagrees with an aggressive ML label, stages and confidence are corrected rather than silently overwritten.
Pattern and structure outputs participate in support and resistance storytelling—so automated levels connect back to detected geometry rather than arbitrary lines.
The Market Screener ranks and filters using the identical stage-and-signal vocabulary—Weinstein phase, buy/sell/hold, model confidence, SATA, liquidity, relative volume, sector, industry, and pattern facets—with a Trending composite that rewards aligned setups.
Portfolio intelligence, automated agents, and hunt workflows consume the same services—this page is the conceptual spine that ties those experiences together.
Open ML predictions, latest stage analysis, and signal blocks from any ticker—then compare against how the same name ranks in the screener.