@@ -43,13 +43,14 @@ The runtime surfaces four strategy labels. Each has different risk shape, execut
4343| --- | --- | --- | --- |
4444| ` near_zero ` | Paired YES+NO arbitrage on one market | Symmetric two-leg execution, strict gate checks on spread/depth/staleness, strongest fill-discipline posture | ` strategyMode=near_zero_risk ` , `signalMode=near_zero| both` , ` edgeRequired` , ` minPairedFillRate` , ` maxLegSkewMs` |
4545| ` ev ` | Single-sided directional execution from calibrated probability/insight | One-leg intent (`side=yes| no`), confidence/edge thresholding, cooldown + per-market/portfolio EV notional caps | `signalMode=ev| both` , ` evEdgeRequired` , ` evConfidenceMin` , ` evCooldownSeconds` , ` evMaxPerMarketNotional` , ` evMaxPortfolioNotional` |
46- | ` fw_projection ` | Frank-Wolfe optimized projection for dependency-aware opportunity selection | Solver-driven lower-bound checks, only converged/feasible opportunities proceed, dependency graph quality affects candidate set | ` fwDependency* ` , ` fwGapAbsTolerance ` , ` fwGapRelTolerance ` , ` fwMaxLoopRuntimeMs ` , ` fwMinEdgeThreshold ` |
46+ | ` fw_projection ` | Frank-Wolfe optimized projection for dependency-aware opportunity selection | Solver-driven lower-bound checks; strict profiles require converged loops, while permissive profiles may accept approximate positive iterates | ` fwDependency* ` , ` fwGapAbsTolerance ` , ` fwGapRelTolerance ` , ` fwRequireConverged ` , ` fwMaxLoopRuntimeMs ` , ` fwMinEdgeThreshold ` |
4747| ` fw_basket ` | Multi-market FW basket execution | Basket-level selection/ranking, market count bounds, configurable basket execution mode (` sequential_failfast ` or ` batch_best_effort ` ) | ` fwBasketMinMarkets ` , ` fwBasketMaxMarkets ` , ` fwBasketExecutionMode ` , ` fwMaxPerMarketNotional ` , ` fwMaxPortfolioNotional ` |
4848
4949Operator notes:
5050- Dashboard intent labels are normalized to ` near_zero ` , ` ev ` , ` fw_projection ` , and ` fw_basket ` .
5151- ` ev_single_side ` is displayed as ` ev ` .
5252- If explicit strategy metadata is missing, IDs can infer strategy (` :fw: ` , ` :fwb: ` , ` :yes: ` , ` :no: ` ).
53+ - Preset split: ` near_zero ` , ` moderate ` , and ` high ` keep FW convergence strict, while ` extra_high ` leaves ` fwRequireConverged=false ` for exploratory approximate-iterate operation.
5354
5455## Minimum Requirements
5556
@@ -134,6 +135,7 @@ In `.env`:
134135
135136``` bash
136137TRADING_MODE=paper
138+ RISK_PROFILE=high
137139TRADING_ENABLED=true
138140OPS_API_TOKEN=replace-with-secure-token
139141# Optional dev-only localhost token prefill for /ops/* login
@@ -366,7 +368,7 @@ Set up OpenPolyTrader locally in safe paper mode.
366368Requirements:
3673691) Install backend and dashboard dependencies.
3683702) Create .env and dashboard/.env from examples.
369- 3) Set TRADING_MODE=paper, TRADING_ENABLED=true.
371+ 3) Set ` TRADING_MODE=paper`, ` TRADING_ENABLED=true`, and `RISK_PROFILE=high` .
3703724) Set OPS_API_TOKEN and use runtime ops session login on /ops/*.
3713735) Start with `npm run paper:up` (alias: `npm run dev:ops`).
3723746) Verify:
@@ -409,7 +411,6 @@ curl -H "Authorization: Bearer $OPS_API_TOKEN" http://localhost:3000/health
409411
410412- API reference: ` docs/API.md `
411413- Architecture (with end-to-end event flow diagrams): ` docs/ARCHITECTURE.md `
412- - Operator strategy report: ` docs/Operations/operator-strategy-report.md `
413414- Local setup spec and quickstart details: ` docs/Development/setup.md `
414415- Full command reference (start/help/stop/kill + diagnostics): ` docs/Development/commands.md `
415416- Environment variable descriptions and defaults: ` docs/Operations/environment-reference.md `
0 commit comments