Institutional conviction
Is this a genuine, new, aggressive opening position? It weights the two hardest things to fake: that the position is new, and that it was taken aggressively.
A repeatable, evidence-driven process that converts market structure and alternative data into investable signals — powered by SENTINEL, our dual-engine framework for detecting institutional positioning and informed flow.
A repeatable process converts market structure, alternative data, and model evidence into investable signals. Hypotheses are framed by domain experts, then tested at scale by researchers.
Low-latency routing, cost analytics, and capacity constraints are modeled before capital is deployed. Every signal is validated against realistic transaction costs.
Signals earn capital through drawdown behavior, regime sensitivity, and portfolio-level contribution—not backtests alone. The committee gates deployment on all three.
Production systems monitor how each signal behaves under real capital constraints, feeding live evidence back into research for the next cycle.
A fixed sequence that removes garbage cheaply before any expensive scoring — the same discipline that keeps our surfaced list small and clean.
Pull the full universe with size, price, bid/ask, exchange, and underlying context from the live intraday feed.
Collapse one order split across venues into a single logical print; net spread legs and stock hedges so one directional-looking leg is never mistaken for conviction.
Drop closing trades, liquidations, neutral rolls, and earnings IV-crush traps before scoring. Noise is removed before it can mislead.
Independent engines grade conviction and direction separately, each on transparent, inspectable components rather than one opaque number.
A signal is surfaced only when both engines agree past a strict gate. Below the gate it becomes a watch item, never a conclusion.
Every candidate is graded on two questions that must agree — the same logic that powers SENTINEL.
Is this a genuine, new, aggressive opening position? It weights the two hardest things to fake: that the position is new, and that it was taken aggressively.
Is the direction real and likely to survive — not a hedge, aligned with trend and catalyst? It catches positions that are really volatility trades with no directional view.
A print is shown only when both engines clear the bar at the same time. The overlap of high conviction and real direction is the high-conviction zone.
Our dual-engine framework for detecting institutional options positioning and informed flow — the analytical core of how conviction is scored and surfaced.
Share a mandate, allocator question, or partnership inquiry with the investment office.