1. Provided Data Files & Candidate API Key
AUTHORIZED CREDENTIALStq_sandbox_usr_9941a7bc81e3d042
Header: X-TQ-KEY: tq_sandbox_usr_9941a7bc81e3d042
2. Detailed Strategy Description & Math Formulation
MILLENNIUM STATARB // TIER 13. Strategy Entry & Exit Signals Brief
SIGNAL LOGIC SPECIFICATIONYour quantitative engine must continuously evaluate the incoming tick buffer and execute the following 4 deterministic trading states:
Long Spread Under-Valuation
When the dynamic spread drops ≥ 2.0 standard deviations below mean, Asset A is undervalued relative to Asset B. Send Immediate-Or-Cancel (IOC) buy order for Asset A and sell order for β fraction of Asset B.
Short Spread Over-Valuation
When the dynamic spread rises ≥ 2.0 standard deviations above mean, Asset A is overvalued relative to Asset B. Send IOC sell order for Asset A and buy order for β fraction of Asset B.
Equilibrium Reversion Target
When the spread converges back to within 0.35 standard deviations of the Kalman center line, close both leg positions simultaneously to crystallize net arbitrage profit.
Hard Risk Circuit Breaker
If spread diverges beyond 3.80 standard deviations (structural regime breakdown) or portfolio trailing drawdown breaches 3.0%, trigger emergency flat and halt execution.
4. Strategy Sample Signal Chart with Entry & Exit Markers
VISUAL EXECUTION BLUEPRINTStats You Must Code & Match:
Your strategy implementation in the sandbox will be backtested against the provided dataset. Your code output must match our desk ground truth within ±2.50% tolerance.
Required Target Performance
Solving Protocol Roadmap:
Quantitative System Architecture Blueprint
Click any modular pipeline node below to inspect its input/output interface, mathematical formulation, and failover strategy.
01. L2 Market Data Ingestion Feed
Binance WebSocket depth20@100ms • Zero-copy ring buffer parsing
02. Cointegration & Residual Engine
Engle-Granger Two-Step • Augmented Dickey-Fuller stationarity check
03. Adaptive Kalman State Estimator
Recursive 2D state-space model • Dynamic βt slope & innovation error
04. Normalized Z-Score Signal Generator
Rolling Welford variance • Entry threshold |Z| ≥ 2.0 • Mean exit |Z| ≤ 0.35
05. Millennium Mandate Drawdown Guard
3.0% trailing drawdown circuit breaker • Half-Kelly position sizing
06. Smart Order Router & L2 Book Matcher
Immediate-or-Cancel limit order routing • Queue position & slippage penalty
03. Adaptive Kalman State Estimator
K_t = P_{t|t-1} F_t^T (F_t P_{t|t-1} F_t^T + R)^{-1}; \theta_t = \theta_{t|t-1} + K_t e_t
def step(self, px_a, px_b): ...
Backtest Performance Concurrence Audit
Dataset: 5,000 High-Frequency Ticks • Capital: $100,000 USD| Performance Metric | Desk Benchmark Target | Your Strategy Result | Variance (Δ) | Audit Status |
|---|---|---|---|---|
| Cumulative Net PnL | +18.42% (+$18,420) | +18.15% (+$18,150) | Δ -0.27% | MATCHED (98.5%) |
| Annualized Sharpe Ratio | 3.92 | 3.89 | Δ -0.03 | MATCHED (99.2%) |
| Max Trailing Drawdown | -1.42% (≤ 3.0% Guard) | -1.45% | Δ -0.03% | COMPLIANT (< 3.0%) |
| Win Rate | 76.4% | 75.9% | Δ -0.5% | MATCHED (99.3%) |
| Total Pair Fills | 55 Round-Trips | 54 Round-Trips | Δ -1 Fill | MATCHED (98.2%) |
| Mean Execution Latency | ≤ 0.25 ms | 0.18 ms | Δ -0.07 ms | PASSED (FASTER) |
| Millennium Drawdown Invariant | 0 Hard Violations | 0 Violations | 0 | 100% COMPLIANT |
SYSTEM ARCHITECT COMPETENCY AUDIT: BENCHMARK MATCHED (98.9%)
Your end-to-end HFT StatArb engine meets all institutional requirements. Provided API ticks parsed cleanly, Kalman filter state dynamics converged within 38 ticks, and backtest results matched the Millennium desk ground truth with zero drawdown breaches.