A 2-minute execution algorithm for Nasdaq-100 futures. Designed for high-frequency structural edges with strict risk-first position sizing.
2-minute execution · MNQ / NQ futures · Five-contract scale-out
Twelve consecutive months of Nasdaq-100 futures data on two-minute execution bars, tested with a calibrated slippage model (RTH 0.50pt, ETH 1.00pt, news event wideners). The period captures all twelve calendar months of a single trading year — pre-market, regular session, and overnight flow included.
Every entry fires with five contracts split across four roles. C1 (1 contract) is the canary — an early, time-bounded exit that validates direction and locks in base-rate profit. C2 (1 contract) targets the nearest structural swing. C3 (2 contracts) is the delayed runner with an ATR trailing stop. C4 (1 contract) is the extended runner for longer-timeframe moves. The split exists because no single exit style captures the full distribution of winning trades — each role handles a different part of the move.
Every candidate trade must pass two non-negotiable gates before the entry logic considers it. Gate 1: combined signal score ≥ 0.75 (eliminates low-conviction noise). Gate 2: stop distance between 22 and 50 points (tight enough to cap tail risk, wide enough to survive the 15–20pt fakeouts the Nasdaq regularly prints). These gates are enforced in code, not in configuration — they are architectural, not tunable.
Directional bias is computed from a three-timeframe consensus (10-minute, 30-minute, two-hour bars) and gated at a strength threshold of 0.67. No higher-timeframe alignment, no trade. The consensus was validated via a 526-configuration matrix sweep that ranked this combination first by out-of-sample performance.
The C3 runner only stays in the market if C1 exited profitably. If the canary dies, C3 dies with it — immediately. Across 396 backtested trades in a validation run, this single mechanical rule saved $38,430 and cut maximum drawdown from 8.62% to 1.60%. It is the largest documented edge in the system. The principle: letting winners run requires the discipline to kill them fast when the early signal fails.
When two or more order flow signals align with an average strength of at least 0.65, the higher-timeframe hard gate softens to a small score penalty instead of outright blocking the trade. Overridden trades execute at reduced size (C1 + C2 only, two contracts, no runners). The rationale: order flow leads price by 100–300 points in regime shifts, and higher-timeframe bias lags. Live evidence from March 2026 documented seven short signals blocked by the higher-timeframe gate during a 200-point down move that all would have been profitable — this override is the response.