EMP 1TO2 · Methodology
From a research thesis to a tradeable edge
EMP 1TO2 starts from one of the best-documented effects in market-microstructure research — intraday momentum — and rebuilds it for the way a trader actually executes: a fixed 1:2 bracket, native platform orders, 1-tick simulation and an untouched out-of-sample test.
1 · The thesis
Why intraday moves keep going
Academic work over the last decade has documented that, in index markets, a large move during the morning tends to be followed by more movement in the same direction later in the session. Two mechanisms explain it.
Mechanism one — order flow is split
Large investors rarely execute at once; they work orders across the day to limit market impact. A flow big enough to push price outside its normal range early is often still being executed hours later. Gao, Han, Li & Zhou (2018) showed the market's first half-hour return predicts its last half-hour return.
Mechanism two — dealers hedge with the move
As the session progresses, option market makers who are short gamma must buy into rallies and sell into declines to stay hedged — adding to the move instead of fading it. Baltussen, Da, Lammers & Martens (2021) linked this hedging demand to intraday momentum across dozens of futures markets.
The noise band
Zarattini, Aziz & Barbon (2024) turned the effect into a rule: price is treated as noise while it stays within the distance it normally travels from the open by that minute of the day, measured over recent sessions. Only a break outside that boundary is treated as information — and followed.
A Berkeley Haas working paper (Chuk, 2026) then adapted the idea to Nasdaq-100 futures and asked the question that matters to a trader: does it survive being made executable? It did — positive in every calendar year tested — but the paper's own stress tests showed the weak point: its trailing-stop exits were so sensitive to fills that one index point of slippage removed roughly a fifth of the edge.
The band, formally
σt is the average absolute percentage move from the open at minute t over the previous n sessions; O is today's open and C−1 yesterday's close. A close above Ut is a break up, below Lt a break down. This is the published research definition — EMP 1TO2's own construction, lookback and filters are proprietary.
Simulated price and an illustrative band, for intuition only. EMP 1TO2's band construction, entry timing, filters, stop and target placement are proprietary and not shown.
2 · The engineering
What EMP 1TO2 changes
The research proves a mechanism; it doesn't hand you a product. These are the design decisions that turn the thesis into something you can run every day.
A fixed 1:2 bracket
Instead of a trailing stop that slides with every tick, every trade is placed with its risk defined first and a target at twice that risk. The outcome is known in R before entry, and exits no longer depend on chasing a moving stop.
Risk set by the market
Stop distance is derived from current market conditions rather than a fixed number of points, so a quiet session and a volatile one carry comparable risk in R terms.
Native NinjaTrader orders
Orders are placed and managed inside NinjaTrader as a native bracket — no webhook, no alert relay, no extra hop of latency between signal and order.
Tick-accurate testing
Every trade was simulated on 1-tick data with slippage charged on every eligible fill, so the order of stop and target touches inside each minute is resolved tick by tick.
Full-size NQ, one at a time
One contract, never more than one position open, intraday only, with daily trade limits and session time rules. Scale it down to MNQ for smaller accounts.
Settings stay on our server
The exact parameters never ship inside the file. They are delivered encrypted to your licence at start-up — nothing to tune, nothing to leak, nothing to curve-fit.
Why 1:2 changes the maths
With a 2R target and a 1R stop, a system breaks even at a 33.3% win rate before costs. Commission, slippage and trades closed by time rules pull the realised payoff to 1.78×, which lifts the real break-even to 35.99%.
EMP 1TO2 wins 42.01% of trades — a margin of 6.02 percentage points over its own break-even. That margin, repeated over hundreds of trades a year, is the edge.
Scale 0–60%.
3 · The research process
Built on one period, judged on another
A backtest is only as honest as its process. These are the controls behind the EMP 1TO2 backtest.
- 1
Chronological split
The recorded history (20 Sep 2021 – 27 Oct 2026) is split 70% / 30% by elapsed time. The strategy was developed on the first 70% and frozen before the final 30% was evaluated.
- 2
Out-of-sample result
Out of sample: 461 trades, 42.30% win rate, profit factor 1.271, $73,680 net — against 41.90% and 1.296 in sample.
- 3
Tick data, real costs
1-tick historical data, 2 ticks of slippage on every eligible fill and $5.76 commission per trade, all inside every result shown.
- 4
Execution stress
The same strategy re-run at 1, 2 and 5 ticks of slippage and paired trade by trade, plus an overlay that charges every trade extra adverse ticks directly.
- 5
Every metric recalculated
Every figure in the backtest is recomputed from the raw trade ledger and reconciled against NinjaTrader’s own summary, with conventions disclosed where they differ.
- 6
The bad news, published
The median trade is a loss (-$561). The longest losing run is 13 trades. The deepest drawdown is -$12,687 per NQ contract. You see all of it before you buy.
| In sample (first 70%) | Out of sample (final 30%) | |
|---|---|---|
| Trades | 1,234 | 461 |
| Win rate | 41.90% | 42.30% |
| Profit factor | 1.2961 | 1.2714 |
| Average trade | $134.65 | $159.83 |
| Payoff ratio | 1.797× | 1.734× |
| Net per elapsed year | $46,533 | $48,147 |
| Max. closed-trade drawdown | -$12,687 | -$11,262 |
4 · Sizing it like a researcher
Plan for the drawdown you haven't seen yet
A single historical path understates the tail. Resample the backtest thousands of times, pick your size and a drawdown limit — for example your prop firm's — and see how often it would have been breached.
2,000 simulated 3-year paths of 997 trades each, drawn with replacement from the 1,695 recorded trades and scaled to your size. Each trade is treated as independent, so these paths are deliberately harsher than the single historical sequence — which is the point: the next five years will not replay the last five. Plan around the 95th–99th percentile, not the actual worst.
The research paper reaches the same conclusion from its own bootstrap: size from the stressed drawdown distribution, not from the best historical path, and stay small until a live record justifies more.
5 · Limitations
What this is not
A backtest, not a live record
Every result is a simulation. Live fills, latency and queue position can be worse than any historical fill engine assumes.
Regimes change
The research found intraday momentum in Nasdaq futures strongest since 2020. A market that changes character can weaken or remove the effect.
Selection risk
Any strategy chosen from research carries the risk that its history flatters it. The out-of-sample split reduces that risk; it cannot remove it.
Losing streaks are normal
At a 42% win rate, runs of losses are certain. The backtest contains a 13-trade losing streak and a 110-day drawdown. You must be able to sit through both.
References
- Baltussen, G., Da, Z., Lammers, S., & Martens, M. (2021). Hedging demand and market intraday momentum. Journal of Financial Economics, 142(1), 377–403.
- Chuk, J. (2026). Noise-band intraday momentum in Nasdaq-100 futures. Working paper, Haas School of Business, University of California, Berkeley.
- Gao, L., Han, Y., Li, S. Z., & Zhou, G. (2018). Market intraday momentum. Journal of Financial Economics, 129(2), 394–414.
- Zarattini, C., Aziz, A., & Barbon, A. (2024). Beat the market: an effective intraday momentum strategy for S&P 500 ETF (SPY). Working paper, SSRN.
EMP 1TO2 is an independent implementation inspired by this research; the authors are not affiliated with Emporio Systems and do not endorse it.
The research is public. The implementation isn't.
Licensed, device-locked, with parameters delivered encrypted from our servers.
For education and analysis only. Nothing here is financial advice. Trading futures carries substantial risk of loss. Past and back-tested performance is not indicative of future results. Read the full risk disclaimer.