ORB strategy on ES and NQ: 10 years backtested
An ORB strategy on ES and NQ is easy to write and easy to fool yourself with. The rules need exact times, the costs need to hurt, and ten years of results are only worth reading year by year, with the range length checked on both markets.
By Soren Huang · Published · updated · Editorial policy
Contents
Does an ORB strategy still work on ES and NQ?
The rules themselves are defined in opening range breakout strategy: exact rules and how to backtest it; this page runs them on ES and NQ. An ORB strategy there buys a break of the first 15 minutes after the 9:30 ET open and is flat by the close. Range length and costs decide whether it holds up.
I'll say up front what I think. Most ORB strategy backtests you see online are fitted to one range length on one market, with costs that assume you get filled at the breakout price. That's the version that falls apart.
What are the exact rules?
Here's the base version I'd test on both markets before touching anything. It's deliberately plain.
- Session. Regular trading hours only, 9:30 to 16:00 ET. The overnight Globex session is ignored when drawing the range.
- Range. The high and low of the first 15 minutes, so 9:30 to 9:45 ET.
- Entry. Buy stop 1 tick above the range high, sell stop 1 tick below the range low. One trade per day; whichever side fills first wins, the other order is cancelled.
- Stop. The other side of the range.
- Target. None. The trade rides until the stop or the exit time.
- Exit time. Flat at 15:55 ET, whatever the position is doing.
Two details trip people up. The range has to follow New York time through daylight saving changes, or for a few weeks a year your "first 15 minutes" is really a slice of the pre-open. And the stop must be the far side of the range, not a fixed number of points, because a 15-minute range on NQ in a quiet August is nothing like one in a busy March.
How are ES and NQ different for an ORB?
NQ moves more. It is the same idea on a jumpier index, and that changes the costs and the stops.
ES is $50 a point and trades in 0.25-point ticks, so one tick is $12.50. NQ is $20 a point, also 0.25-point ticks, so one tick is $5. In ticks, an NQ opening range is usually much wider than an ES one. A fixed slippage of 2 ticks per side is a small share of an NQ range and a much bigger share of an ES range.
So the same rules can look worse on ES purely because costs eat a bigger slice of each trade. I wouldn't call that a failure of the ORB idea on ES. It's a cost problem, and you see it in the average trade measured in ticks.
The other difference is the gap. NQ opens away from the previous close more often and by more, so the first 15 minutes can already contain a big part of the day's move. A breakout after that has less room left.
Which ORB settings overfit first?
The range length. Always the range length.
Test 5, 10, 15, 30 and 60 minutes and one of them will look best. On ten years of ES that's normal noise. What I want to see is a plateau: if 15 minutes is good, 10 and 30 should be at least decent. A single spike at 15 with losers either side is a fitted number, not an edge.
Next on my list is the breakout buffer. Moving the entry from 1 tick to 3 ticks beyond the range often "improves" a backtest because it skips the days where price pokes out and snaps back. Maybe it's real. It can flip from one five-year half of the data to the other, though, so I trust it least of all the settings.
After that come the day-of-week filters ("skip Mondays"), the exit time, and a profit target as a multiple of the range. Each one is a new knob, and each knob you add gets fitted to the same ten years. My rule: one filter at most, and it has to make sense on NQ too without retuning.
How conservative should the costs be?
Conservative enough that the backtest annoys you. The open is the most expensive moment of the day to trade with a stop order.
In AlphaPrime (Valynox), the cost model defaults to Conservative. For ES and NQ that means commission at 1.5 times the standard broker estimate, about $3.38 per side, plus one more tick of slippage per side than the standard model, so 2 ticks per side. On ES, one round trip then costs about $50 in slippage plus about $6.76 commission. On NQ it's about $20 plus the same commission.
I got this wrong once. An earlier version of the conservative profile pushed the extra pessimism into the commission field, and it showed up as around $40 per side. No broker charges that, and it hid where the cost really comes from. On 19 October 2026 I changed it so commission stays at a real broker level and the extra cost sits in slippage ticks, where the open actually hurts you.
My threshold: if the average ORB trade after conservative costs is under 2 ticks, I don't trade it. A 1-tick average edge disappears the first week the open gets fast.
What should you read in ten years of results?
Read it year by year, not as one total. A 2016 to 2025 total can be carried by one or two violent years, and an opening range breakout loves violent years. A quiet one can bleed for twelve months straight. The periodic performance in AlphaPrime's performance report splits a ten-year ORB test into years, with the drawdown and trade list beside it.
What I look at, in order:
- Net profit per calendar year, and how many years were negative.
- Average trade in ticks after costs, per market.
- Long side and short side separately. On a rising index the long side often carries everything.
- Trades per year. Roughly one setup a day gives a couple of hundred a year, which is plenty to judge.
- The maximum drawdown in dollars against what you'd actually fund the account with.
Then the part most people skip: hold the last few years out, build on the rest, and see what the held-out years say. Walk-forward optimization does this repeatedly, re-picking the range length on each window and testing it on the next. If the range length the walk-forward keeps choosing jumps around from 5 to 60, there was never a stable setting to find.
For a sense of what held-out data does to a strategy list: on 15 October 2026 I ran AlphaPrime's factory settings on ES 60-minute bars with an RTX 3060 Ti, holding out everything from 2022-10. Of the strategies that passed walk-forward, 91% were profitable on the held-out bars, against 53% of the ones that failed it. That wasn't an ORB test. It's why I don't trust any ORB total that wasn't checked the same way.
Can AlphaPrime search ORB variants for you?
Yes, and that's also the risk. Strategy generation in AlphaPrime can combine entry rules with different filters and exits on ES and NQ and backtest thousands of combinations. The more variants you search, the more one of them will look great by luck.
So I'd use it the other way round: fix the core ORB rules above, let the generator vary only the exit and one filter, keep the conservative cost model on, and send everything through walk-forward before reading a single equity curve.
Start with ES and NQ together, 15-minute range, conservative costs, and write down before you run it what result would make you drop the idea. Then run it.
Frequently asked questions
What is the best opening range length for ES?
There isn't one fixed answer, and a backtest that names one exact length should make you suspicious. Look for a range of lengths, for example 10 to 30 minutes, that all hold up on both ES and NQ.
Should an ORB strategy include the overnight session?
For ES and NQ I build the range from the regular session only, starting at 9:30 ET. Including Globex makes the range much wider and turns it into a different strategy. Test it as a separate idea if you want it.
How much slippage should I charge on an opening range breakout?
At least 2 ticks per side on ES and NQ for stop orders at the open. If the strategy only works with 1 tick or less, it is unlikely to survive real fills.
Is NQ better than ES for an opening range breakout?
NQ ranges are wider in ticks, so fixed tick costs take a smaller share of each trade. That doesn't make it better on its own. Run the same rules on both and trust only what works on both.