Abstract

This paper presents a comprehensive analysis of an intraday momentum breakout strategy for E-mini S&P 500 (ES) and E-mini NASDAQ-100 (NQ) futures. The strategy employs volatility-based "noise area" boundaries — percentile-based rather than standard deviation — to identify momentum opportunities while maintaining strict intraday-only execution. We implement a volatility-targeted position sizing framework with a maximum leverage constraint of 8x and target daily portfolio volatility of 3%. Through rigorous backtesting from 2011–2026 with 1-tick slippage per side, the portfolio achieves 247% total return, 8.5% CAGR, and a Sharpe ratio of 1.18.

Introduction

Intraday momentum strategies exploit short-term price movements that persist over minutes to hours. Unlike daily momentum, intraday approaches eliminate overnight gap risk while leveraging the highly liquid E-mini futures markets ($100B+ daily volume). The strategy targets the regular session window, where momentum effects are empirically strongest, and mandates a flat position by end of day under all circumstances.

Motivation

Three empirical observations motivate the design:

1. Volatility clustering: Intraday price movements exhibit GARCH effects, creating predictable momentum patterns
2. No overnight risk: Intraday-only execution eliminates exposure to overnight gaps — which account for ~35% of buy-and-hold drawdowns in the same period
3. Noise area innovation: Unlike Bollinger Bands (standard deviation), percentile-based boundaries adapt more robustly to regime changes in intraday range distributions

Key Equations

Noise area boundaries (90-day lookback, percentile-based):

where is the daily high-low range.

Signal generation (with 2-bar confirmation and volume filter):

Volatility-targeted position size:

Leverage cap (hard limit at 8×):

Daily loss limit:

Flash crash circuit breaker:

Algorithm Blueprint

Intraday Momentum Breakout — ES & NQ

Inputs: 5-minute OHLCV bars, initial capital $100,000

Outputs: Daily intraday signals, position sizes, mandatory end-of-day flat

Algorithm:

1. PRE-SESSION SETUP (9:30 AM ET)

   • Compute 90-day intraday range history: for each day
   • Upper boundary (opening price plus 75th percentile):
   • Lower boundary (opening price minus 25th percentile):
   • Instrument volatility:

2. SIGNAL DETECTION (9:30 AM - 3:00 PM ET, 5-min bars)

   • For each 5-min bar:
     • If close $> UB$ AND volume AND no open position:
       • Confirm over bars → LONG signal generated
     • If close $< LB$ AND volume AND no open position:
       • Confirm over bars → SHORT signal generated

3. POSITION SIZING

   • Target portfolio risk:
   • Leverage cap: where $m$ = multiplier
   • Hard contract cap:

4. RISK MANAGEMENT

   • Daily loss limit: 5000})$
   • If : flatten all positions, halt new entries
   • Circuit breaker: if , flatten and 30-minute trade halt
   • Strategy halt: if drawdown , require manual intervention

5. EXIT HIERARCHY

   1. Momentum failure: price re-enters noise area (minimum 3-bar hold)
   2. Maximum hold: 78 bars (approximately full trading session)
   3. MANDATORY: flatten all positions at 15:45 ET (end-of-day)

6. PORTFOLIO ALLOCATION (across ES and NQ)

   • NQ Momentum component: 50% portfolio weight
   • ES Momentum component: 25% portfolio weight
   • NQ Long-only component: 25% portfolio weight
   • Daily rebalancing if weight drift from target

Results

| Metric | Portfolio | NQ Mom. | ES Mom. | NQ Long | |---|---|---|---|---| | Total Return | 247.3% | 312.5% | 98.7% | 189.4% | | CAGR | 8.5% | 10.1% | 4.7% | 7.2% | | Sharpe Ratio | 1.18 | 1.05 | 0.67 | 0.94 | | Sortino Ratio | 1.67 | 1.52 | 0.91 | 1.38 | | Max Drawdown | -16.4% | -21.1% | -15.2% | -16.3% | | Win Rate | 54.3% | 52.8% | 50.1% | 61.2% | | Profit Factor | 1.64 | 1.58 | 1.32 | 1.71 | | Total Trades | 3,847 | 2,134 | 1,713 | 412 |

Slippage sensitivity: each additional 0.5 tick reduces CAGR by ~2.7% and Sharpe by ~0.3. Walk-forward out-of-sample Sharpe: 0.94 (20% degradation vs in-sample — within acceptable threshold). COVID 2020 drawdown: -16.4% portfolio peak-to-trough, with volatility targeting automatically reducing average leverage to 1.8× during the crash vs 3.5× normal.

Contributions

• Percentile-based noise area framework replacing standard deviation bands, providing more robust regime adaptation
• Volatility-targeted position sizing with 8× hard leverage cap, demonstrated to automatically de-risk during crisis periods (COVID 2020)
• Walk-forward validation over 15 years (2011–2026) spanning multiple regimes with 20% acceptable out-of-sample degradation
• Quantification of the overnight risk elimination benefit: strategy avoids 100% of gap events that account for ~35% of buy-and-hold drawdowns

Paper

Author

Frankline Misango Oyolo Quantitative Research Division, Arithmax Research Frankline@arithmax.com — Published: March 2, 2026