The gap fill strategy rests on one of the most repeated lines in day trading: "gaps always fill." We tested that claim on Nasdaq-100 futures (NQ) across 3,884 regular sessions from June 2010 to April 2026, fading every opening gap back toward the prior close. The result is a lot less comfortable than the slogan.
What an opening gap is, and why traders fade it
An opening gap is the difference between today's regular-session open (9:30 am ET) and yesterday's regular-session close (4:00 pm ET). Because NQ trades nearly 24 hours, the "gap" is really the overnight move. A gap fade bets that the overnight move was overdone and price will return to yesterday's close during the day session. When it gets there, the gap has "filled."
The popular version of this idea says small gaps are the safest to fade because they fill most often. That's the claim we wanted to test with real NQ gap fill statistics, after costs.
The exact gap fill rules we tested
- Data: NQ continuous front-month futures, 15-minute bars, regular session 9:30 am–4:00 pm ET, June 7, 2010 to April 16, 2026.
- Gap size: (today's 9:30 open − prior session's 4:00 pm close) ÷ prior close, in percent. We sorted trades into four buckets by absolute size: 0.1–0.3%, 0.3–0.6%, 0.6–1.0%, and 1.0% or more. Gaps under 0.1% were skipped.
- Entry: at the 9:30 open, against the gap. Short gap-ups, buy gap-downs.
- Target: the prior day's close (the gap fill).
- Stop: one full gap-length beyond the entry. A 40-point gap-up is shorted at the open with a stop 40 points higher.
- Time exit: if neither is hit, exit at the 4:00 pm close.
- Conservative fill rule: if a single 15-minute bar touches both the stop and the target, we count the stop.
- Costs: 1.0 NQ point per round trip deducted from every trade. Results are NQ points per 1 contract; MNQ = $2/point, NQ = $20/point.
Gap fill statistics by gap size
| Variant | Trades | Win rate | Avg pts | Total pts | Profit factor | Closed-trade DD (pts) | Pts 2010–19 | Pts 2020–26 | PF 2020–26 |
|---|---|---|---|---|---|---|---|---|---|
| 0.1–0.3% | 1032 | 48.1% | -1.15 | -1,183.8 | 0.87 | -1,325.0 | -561.0 | -622.8 | 0.90 |
| 0.3–0.6% | 1059 | 50.0% | -0.92 | -979.0 | 0.95 | -1,623.5 | -786.0 | -193.0 | 0.99 |
| 0.6–1.0% | 616 | 51.0% | 3.61 | 2,226.0 | 1.12 | -983.8 | -22.8 | 2,248.8 | 1.16 |
| 1.0%+ | 529 | 42.9% | -4.60 | -2,432.8 | 0.91 | -6,192.0 | -221.2 | -2,211.5 | 0.91 |
That's 3,236 trades in total. Win rate here means the trade closed in profit, either by hitting the target or by being ahead at 4:00 pm.
The "gaps always fill" myth
Three of the four buckets lost money:
- Small gaps (0.1–0.3%) were the worst per trade: 1,032 trades, a 48.1% win rate, a profit factor of 0.87 and -1,183.8 points ($-2,368 on one MNQ). They lost in both decades.
- 0.3–0.6% gaps were close to breakeven before costs and lost -979.0 points after them.
- 1.0%+ gaps were the most dangerous. They lost -2,432.8 points with a closed-trade drawdown of 6,192.0 points. Big gaps are often news-driven and keep going. 2020 alone cost -2,212.0 points. The last two years flipped positive (2025: 1,826.8; 2026 through April: 1,252.5), which is a good example of how a short streak can make a bad idea look good.
Only the 0.6–1.0% bucket was positive: 616 trades, 51.0% winners, profit factor 1.12, 2,226.0 points ($4,452 on MNQ). Even that comes with a catch.
The robustness check: gap fade before and after 2020
NQ went from about 1,800 to about 26,000 over the test, so recent years dominate point totals. Splitting the positive bucket by decade: the 0.6–1.0% gap fade made -22.8 points in 2010–2019 and 2,248.8 points in 2020–2026 (profit factor 1.16). In other words, it was flat for ten years and all of the profit came after 2020. That doesn't make it worthless, but a profit factor near 1.1 that only showed up in one regime is a thin edge.
Why small gaps did so badly
Two reasons. First, costs: the average winning small-gap trade was only 16.54 points, so a 1-point cost takes a real bite. Second, our conservative rule. A 0.2% gap on NQ can fill and hit a same-size stop inside the same 15-minute bar. When that happens we count the loss. A tick-by-tick test might score some of those as wins, so the small-gap numbers are likely a bit pessimistic. Still, that same noise is what makes small gaps hard to trade with tight stops in real life.
What this backtest does not model
- Fills at the exact open, target and stop prices. Real market orders at 9:30 slip, especially on big gap days.
- Intrabar order of events (handled conservatively, stop first).
- Overnight gap risk is irrelevant here since all trades close by 4:00 pm, but news during the session is not filtered.
- Contract roll days and taxes.
This is a historical backtest. Past results don't guarantee future results, and nothing here is financial advice.
Key takeaways
- Across 3,236 NQ gap fade trades (2010–2026), three of four gap-size buckets lost money after 1 point of costs.
- Small gaps (0.1–0.3%) had the lowest profit factor (0.87) and lost in both decades.
- 1.0%+ gaps had the deepest closed-trade drawdown (6,192.0 points).
- Only 0.6–1.0% gaps were positive (PF 1.12), and nearly all of that came after 2020.
- "Gaps always fill" is not a trading rule. A gap fill strategy needs a stop, and the stop is what breaks it.
Put the research to work on your Desk
Explore the tools behind the research, compare the available access, or open your Desk if you already have an account.
FAQ
Do gaps always fill on Nasdaq futures?
No, at least not in a way you can trade. With a stop of one gap-length and an exit at the 4:00 pm close, fading NQ opening gaps lost money in three of four size buckets from 2010 to 2026 after 1 point of costs.
Which gap size is best to fade?
In our test only 0.6–1.0% gaps were profitable: 616 trades, profit factor 1.12. Almost all of that profit (2,248.8 points) came in 2020–2026, and 2010–2019 was roughly flat (-22.8 points).
Why did small gaps lose money?
Small gaps have small average wins, so a 1-point cost matters a lot. Also, a small gap can fill and stop out inside the same 15-minute bar, and our test counts the stop first in that case, which is conservative.
How is the gap measured?
From the prior regular-session close (4:00 pm ET) to today's 9:30 am ET open, as a percentage of the prior close.
Loomi.Ai provides educational market research, not financial advice. Trading futures and options involves substantial risk of loss, and backtested results are historical and do not guarantee future returns.
Sources
- Loomi backtest: blog_strategy_tests.py + results JSON, Loomi research