What changed
v9 adds the missing institutional layers: Execution Intelligence, Portfolio Brain, trade lifecycle management, edge decay detection, strategy tournament allocation, TUCE copilot routing, and deep integration verification.
Hot path policy
Keep the hot path deterministic: ingest → features → signal → risk → route → order. Use copilots only for explanation, experiments, incident review, and operator guidance.
Operator shortcuts
Shift+P paper mode, Shift+S shadow mode, Shift+L live mode, Shift+X kill switch.
Execution Intelligence Layer
| Capability | Purpose |
|---|---|
| Venue scoring | Choose venue by latency, fills, slip, fees, rejects, stability |
| Algo selection | TWAP/VWAP/POV/Iceberg/IS/Post-only by urgency and size |
| Route memory | Remember venue quality by symbol/session/regime |
| Reaction lab | Estimate self-induced impact and decide execute/probe/wait |
Portfolio Brain
Institutional portfolio controls sit here: concentration, correlation pressure, VaR/ES, gross/net budgets, drawdown governors, hedge hints, and strategy tournament allocations.
Evidence Lab
Backtest, replay, walk-forward, OOS benchmark, and TCA/benchmark stores belong here.
Governance
Promotion gates, incidents, evidence counts, stale-review holds, and mode transitions.
Euler Edge
Includes the five advanced Euler edge modes plus Max Edge Fusion, 64-pattern priors, and strategy compatibility overlays.
TUCE Copilot
Routes long-context synthesis to Gemini-style paths and code/tooling/operator work to Claude-style paths, while keeping trade decisions governed and backend-owned.