Opening Range Breakout Backtest: The 5-Minute ORB on NQ Futures, 2020 to 2026

00The verdict in three lines

• We tested the 5-minute opening range breakout of Zarattini and Aziz (2023): take the direction of the first 5-minute candle of the US session, enter at 09:35, stop at the other side of that candle, target ten times the risk or exit at the close. And the classic version most traders mean by ORB: a stop order one tick beyond the high or low of the same range. One NQ contract, 2020 to 2026, $15.76 a round trip.
• The paper's numbers held, and the rule made money: 24.3% winners and +0.107 R a trade before costs (the paper: about 24% and +0.13 R), +$70,146 on one contract over 1,729 trades.
• It is still buried, like the classic version: the same trades with the direction taken from other days also made money, the rule sat at the 79th percentile of them, without its 10 best trades it loses, and 2026 so far is a losing year for each of the four versions.

01What we tested

The opening range: the five 1-minute bars from 09:30 to 09:35 New York. Its median size here was 51 points ($1,010 on one contract).

• ZA, the paper's rule (Zarattini and Aziz, “Can Day Trading Really Be Profitable?”, SSRN 2023, tested on QQQ from 2016 to 2023): if the first 5-minute candle closed up, buy; down, sell; unchanged, no trade (4 days in 1,733). Entry at the open of the minute starting 09:35; stop at the other side of the first candle. The paper also shows versions with a stop set from the ATR and other targets, chosen after looking at results; we tested the base rule.
• BO, the classic breakout: from 09:35, a buy stop one tick above the range high and a sell stop one tick below the range low; the first one touched is the trade, until 13:30 New York. The stop is one tick beyond the other side of the range.
• Exit, both rules: a target of 10 times the risk (10 R), otherwise the close of the session (16:00, or the exchange's earlier close). Inside a minute that touches both the stop and the target, the stop is taken. One contract, one trade a day.
• Four variants: ORB-ZA-LS, ORB-ZA-L, ORB-BO-LS, ORB-BO-L (LS = long and short; L = long only, the buy trades of the same rule). Four variants of a strategy not tested before on the site, so the development t-statistic bar is 2.0.

Before this test: in September 2026, without a protocol and unpublished, we had looked at 5, 15 and 30-minute opening range breakouts on ES (no proven edge), and the site's owner had tested their own ORB variant on ES. No opening-range number on NQ had been computed. A third-party replication of the paper's rule on index CFDs reproduced the gross result on the Nasdaq-100 and found it near zero after CFD costs (MQL5 blog); futures costs are much smaller, which is one reason to test it here.

02Data and costs

E-mini Nasdaq-100 futures (NQ), NinjaTrader's back-adjusted 1-minute history, January 2020 to 25 September 2026 (back-adjustment shifts prices by constants at rolls, which changes no move inside a day). Periods: development 2020 to 2023, validation January 2024 to March 2025, holdout April 2025 to September 2026, already used by our earlier NQ tests. Costs: $15.76 a round trip ($5.76 commission plus one tick of slippage on each fill), $20 a point. Engine check: hand-built days (up, down and unchanged first candles, a stop and a target in one minute, gaps, an early close, a missing minute, every breakout case), the exchange calendar written out twice, and all four variants matched trade by trade against an independent re-implementation (orbq_selftest_result.txt). In the first run one hand-made check was written wrong in the test itself (it expected a target the case never reaches); it was corrected before any result was computed.

03Results

Dollars per contract. The 95% range is the Wilson interval for the win rate after costs.

Cumulative net dollars per contract by month. * Holdout already used before this test.

By period

All four made money over the whole sample, and none came close to the development bar (t of 2.0). The paper's rule lost $15,750 in the holdout; the classic breakout made a little.

04The paper's numbers, on futures

The paper reports about 24% winning trades and +0.13 R a trade on QQQ. On NQ futures the same rule had 24.3% winners and +0.107 R a trade before costs, close to both. The cost of a round trip, $15.76, is only 0.028 R next to an opening range worth about $1,010: on futures, unlike on the CFDs of the third-party replication, costs are not what stops this rule. The 10 R target was reached in only 1.9% of trades; 74.9% were stopped and the rest closed at the session close.

05Where the money came from

The three best trades of the paper's rule: 24 February 2022 (long, +$15,509, held to the close); 9 April 2025 (long, +$14,684, the 10 R target); 13 October 2022 (long, +$11,034, the 10 R target). The ten best together made +$105,262, more than the rule made in total:

Year by year (net dollars, trades)

2026 so far is a losing year for each of the four versions; the paper's rule is down $28,193. A rule that lives on rare trend days can go a long time without one.

06Is it the first candle or the trade's shape?

The benchmark is a shifted-direction control: the series of daily directions (buy, sell or none) is moved by a random offset of at least 5% of the days and applied to the real days, with the same opening range, stop, target, exit and costs (for the breakout, only the moved side's order is placed). 500 runs with fixed seeds; the bar is the 95th percentile of the average net dollars per trade.

The control made money in every variant (a median of +$18.66 a trade for the paper's rule): with the stop at the edge of a 5-minute range and the exit at the close, the trade pays on trend days whichever way it was pointed. The first candle's direction added something (the 79th percentile), not enough to tell it from chance.

07The gates

08Verdict by variant

Fragile, in our protocol's sense: positive overall, but not without its ten best trades. All four also failed the development t-statistic and the control. Not tested: the years before 2020, the paper's optimised versions, other opening ranges, other markets.

TWhat this shows, and what to take into your own tests

What the results show

• On NQ futures from 2020 to 2026 the paper's own figures came back: 24.3% of the trades made money before costs (the paper: about 24%) and the average trade made +0.107 R before costs (the paper: about +0.13 R). Costs on futures are small next to an opening range, 0.028 R a trade.
• The rule still made +$70,146 on one contract over 1,729 trades, but the first candle's direction added little: the same trades with the direction taken from other days made +$18.66 a trade on average, and the rule sat at the 79th percentile of them (95 needed).
• The profit rests on a handful of trend days: without its 10 best trades the paper's rule loses $35,116, and 2026 so far is a losing year for each of the four versions (from -$8,135 to -$32,979).

What remains unknown

• The years before 2020 and the QQQ data of the paper. Our futures file starts in 2020.
• The paper's optimised versions (a stop set from the ATR, other targets), which were chosen after looking at results. We tested the base rule only.
• Other opening ranges (15 or 30 minutes) and other markets: one market per article.

Why it was buried

The shape of the trade does most of the work: a stop at the other side of a 5-minute range, and no exit until the close unless a 10 R target is hit. On the few days when NQ trends hard all day, that shape pays many times the risk, whichever way it points; on most days it loses a small R. Our control kept that shape and took the direction from other days, and it made money too. So the result is a bet on rare trend days, not on the first five minutes, and with the profit resting on about ten days, the protocol cannot call it an edge.

What you can apply to your own tests

• When a paper reports a big return, test the base rule on data after or outside the paper, with real costs. Here the base numbers held, which is rarer than it sounds.
• Then ask what part of the result the signal explains. Keep the exits and randomise the entry or the direction: if that makes money too, the exits are the strategy.
• Remove the ten best trades. A strategy that lives on a few trend days can be right for years and still lose a year like 2026.

A note on the numbers: drawdown figures are measured on closed trades unless a table says it includes open losses. Open losses inside a trade are not counted in a closed-trade drawdown, so real account risk is higher, especially for systems with no stop. Backtests are not live trading.

FAQCommon questions

Does the 5-minute ORB strategy work?

On NQ futures it made money but did not pass our gates. The paper's rule (the direction of the first 5-minute candle, in at 09:35, stop at the other side, 10 R target or the close) made +$70,146 on one contract over 1,729 trades from 2020 to 2026, after costs. Without its 10 best trades it lost $35,116, and it sat at the 79th percentile of a control with random directions (95 needed).

What is the opening range breakout strategy?

You mark the high and low of the first minutes of the session (here 09:30 to 09:35 New York) and trade in the direction the price leaves that range, with a stop at the other side. The version of Zarattini and Aziz (2023) takes the direction of the first 5-minute candle and enters at 09:35; the classic version waits for a break of the high or the low.

Is the Can Day Trading Really Be Profitable paper right?

Its base numbers came back on NQ futures: about 24.3% winning trades and +0.107 R a trade before costs, against about 24% and +0.13 R in the paper. But in our test the profit came from the trade's shape on rare trend days rather than from the first candle: random directions with the same stop and exit made +$19 a trade on average too.

Does the classic opening range breakout work on NQ?

Not proven either. Buying a break above the 5-minute high and selling a break below the low, with the same stop, target and exit, made +$93,733 on one contract from 2020 to 2026 and passed three of our six gates, but it was no better than random directions and lived on its best trades.

How was the opening range breakout tested?

On NQ futures 1-minute data from 2020 to 2026, one contract, $15.76 of costs a trade, with rules written and dated before the data was run, costs included on every trade, results split into development, validation and holdout periods, a random-entry control as the benchmark, and every variant reported. Read the protocol.

09Check our arithmetic

Code provenance: our own code, written from the rule described in the paper and the classic breakout fixed in the specification. No third-party source code is used or published.

10Data notes

The NQ file has known holes (data_holes.json). Days missing a minute of the opening range or the 09:35 minute have no trade; removing the trades whose day touches a known hole changes no verdict:

11Corrections

17 Oct 2026, wording: three sentences that said every version lost money in 2026 now say that 2026 so far is a losing year for each version (over the whole test all four made money, as the tables show), and our earlier unpublished ES test is described as finding no proven edge rather than no edge. No number or result changed. 22 Oct 2026, navigation: added a link to the new guide comparing trading costs across all our tests. No result changed. 22 Oct 2026, navigation: added a link to the new guide comparing trend following and mean reversion across all our tests. No result changed.

+More tests and guides

• Noise area breakout nq
• Intraday momentum strategy nq
• Vwap strategy backtest
• Fibonacci scalping strategy
• Trend following vs mean reversion: this test next to the others of its style
• Free prop firm challenge calculator
• What all the backtests found so far
• The backtesting protocol
• How to backtest a trading strategy without fooling yourself
• Backtesting glossary