Study · Original data
Opening range breakouts: do they follow through? A placebo test
Measured in units of its own range, an opening-range breakout was no more likely to keep running than the same breakout at a random time of day: after the New York open, crypto reached one more 15-minute range width 62-67 % of the time, against 75-77 % for the placebo.
Key findings
- No edge in range units: across 9 markets, 3 opens and 5 range lengths (135 combinations), breakouts at the real open reached 1x their own range width significantly less often than placebo breakouts in 50 cases and more often in only 2.
- In crypto the New York open is clearly worse than placebo and London is level: with a 15-minute range, 2x the range was reached 35-40 % of the time after a New York breakout vs 53-56 % at random times; after a London breakout 1x was reached 75-78 % vs 75-77 %.
- Most first breakouts do not hold: in crypto and gold, 69-77 % of breakouts of the 15-minute range closed back inside it within 15 minutes - about the same as at random times (73-80 %).
- Measured against the hour before the open (the pre-market balance), the open does move more: in crypto, 1x that hour's range was reached 57-60 % of the time after a London breakout vs 51-52 % for placebo, and about half of 108 combinations were significantly above placebo (7 below). This mostly reflects the open being more volatile than the quiet hour before it - in New York the first 15 minutes were 0.76-0.91x the pre-market range vs 0.45-0.46x at random times - not a sign that the breakout direction predicts anything.
- Gold over three years: Tokyo and London breakouts beat placebo in pre-market units (15-minute range: 68 % vs 57 % and 59 % vs 50 %), but New York was below it (40 % vs 50 %; difference -9.5 points, 95 % CI -14.7 to -4.3). Silver's New York open was weaker still (38 % vs 56 %, 10 months of data).
- The US 500 New York open showed the largest gap in pre-market units (68 % vs 43 % for 1x and 43 % vs 19 % for 2x, 15-minute range), but it rests on 148 sessions over 7 months, and in range units the same breakouts were below placebo (66 % vs 79 %). Treat it as unconfirmed.
Breakouts against the placebo (15-minute opening range)
| Market | Open | Days | Reached 1× range real / placebo | Reached 2× range real / placebo | Reached 1× pre-market hour real / placebo | Back inside within 15 min real / placebo |
|---|---|---|---|---|---|---|
| Bitcoin | Tokyo | 780 | 76% / 78% | 53% / 59% | 58% / 53% | 77% / 78% |
| Bitcoin | London | 782 | 77% / 75% | 56% / 54% | 59% / 51% | 72% / 77% |
| Bitcoin | New York | 782 | 62% / 77% | 35% / 56% | 59% / 52% | 73% / 75% |
| Ethereum | Tokyo | 780 | 77% / 77% | 54% / 56% | 58% / 52% | 76% / 80% |
| Ethereum | London | 782 | 78% / 76% | 54% / 54% | 59% / 51% | 73% / 78% |
| Ethereum | New York | 782 | 62% / 77% | 39% / 55% | 54% / 52% | 73% / 76% |
| Solana | Tokyo | 780 | 75% / 76% | 51% / 56% | 61% / 51% | 76% / 78% |
| Solana | London | 782 | 75% / 75% | 56% / 53% | 57% / 50% | 73% / 78% |
| Solana | New York | 782 | 65% / 75% | 40% / 55% | 56% / 54% | 71% / 78% |
| XRP | Tokyo | 780 | 76% / 79% | 52% / 60% | 63% / 56% | 74% / 77% |
| XRP | London | 782 | 78% / 77% | 57% / 54% | 60% / 52% | 69% / 75% |
| XRP | New York | 782 | 68% / 76% | 40% / 53% | 56% / 49% | 76% / 75% |
| Gold | Tokyo | 757 | 80% / 79% | 60% / 59% | 68% / 57% | 76% / 75% |
| Gold | London | 765 | 77% / 77% | 54% / 55% | 59% / 50% | 75% / 73% |
| Gold | New York | 766 | 58% / 74% | 31% / 55% | 40% / 50% | 74% / 77% |
| Silver | Tokyo | 210 | 75% / 77% | 55% / 54% | 67% / 52% | 74% / 74% |
| Silver | London | 211 | 74% / 74% | 49% / 57% | 53% / 53% | 77% / 81% |
| Silver | New York | 212 | 54% / 80% | 30% / 58% | 38% / 56% | 77% / 73% |
| US 500 | Tokyo | 147 | 63% / 77% | 32% / 58% | 57% / 51% | 76% / 72% |
| US 500 | London | 148 | 80% / 83% | 49% / 63% | 60% / 57% | 78% / 61% |
| US 500 | New York | 148 | 66% / 79% | 36% / 52% | 68% / 43% | 70% / 66% |
| US 100 | Tokyo | 147 | 63% / 80% | 32% / 65% | 62% / 63% | 74% / 77% |
| US 100 | London | 148 | 74% / 78% | 48% / 58% | 60% / 53% | 76% / 72% |
| US 100 | New York | 148 | 46% / 75% | 16% / 56% | 57% / 51% | 78% / 70% |
| WTI oil | Tokyo | 150 | 72% / 81% | 47% / 64% | 72% / 61% | 75% / 66% |
| WTI oil | London | 151 | 78% / 76% | 52% / 54% | 63% / 50% | 68% / 74% |
| WTI oil | New York | 152 | 75% / 78% | 51% / 56% | 51% / 49% | 77% / 72% |
Data per market
| Market | Instrument | Period | Weekdays |
|---|---|---|---|
| Bitcoin | perpetual futures | 2023-11-01 to 2026-10-30 | 782 |
| Ethereum | perpetual futures | 2023-11-01 to 2026-10-30 | 782 |
| Solana | perpetual futures | 2023-11-01 to 2026-10-30 | 782 |
| XRP | perpetual futures | 2023-11-01 to 2026-10-30 | 782 |
| Gold | spot price | 2023-11-01 to 2026-10-30 | 766 |
| Silver | perpetual contract | 2026-01-07 to 2026-10-30 | 212 |
| US 500 | perpetual contract tracking the index | 2026-04-07 to 2026-10-30 | 148 |
| US 100 | perpetual contract tracking the index | 2026-04-07 to 2026-10-30 | 148 |
| WTI oil | perpetual contract | 2026-04-01 to 2026-10-30 | 152 |
How it was measured
We used 1-minute price data for perpetual futures on Bitcoin, Ethereum, Solana and XRP and for spot gold from November 2023 to October 2026, plus shorter series (7-10 months up to October 2026) of perpetual contracts tracking silver, WTI oil, the US 500 and the US 100. For every weekday we marked three opens - Tokyo 09:00, London 08:00 and New York 09:30 local time, with daylight saving handled. The opening range is the high and low of the first 5, 15 or 30 minutes after the open; the pre-market balance is the high and low of the 60 minutes before it. A breakout is the first 1-minute candle that closes above the range high or below the range low once the range is complete. We then checked whether price traveled 1x and 2x the range width - and, separately, 1x and 2x the pre-market width - beyond the broken edge before the next session opened (New York: six hours), counting only targets that were still ahead of the breakout candle's close, and whether it got there before touching the opposite edge of the range. As a placebo, exactly the same rules were applied at a random time on the same day, at least two hours away from all three opens, using that time's own preceding hour and range. If the real opens look like the placebo, the open adds nothing beyond ordinary volatility. Definitions, the placebo and the train/test split were written down before the first result; the final 12 months (the second half for the shorter series) were held back as a test period. Differences are shown with 95 % confidence intervals. The study measures how far price travels after a breakout, not which way it will break; it is not a trading system or advice.
Limits of this study
- Only crypto and gold have three years of data (about 760-780 sessions per open). Silver has about 10 months (210 sessions), and the US 500, US 100 and WTI oil about 7 months (148-152 sessions); their confidence intervals are roughly +/-10 points. A title promising 'three years' must not cover the US 500.
- The pre-registered plan specified three years for metals, oil and indices; only gold got that. The other series were shorter than planned.
- Reaching a distance is not a trade result: no stop, no costs, no direction call and no profit or loss were measured.
- Results in pre-market units are partly mechanical: the hour before an open is usually quiet, so multiples of it are easier to reach once volatility picks up. In units of the opening range itself, the advantage disappears or reverses. The 'clean' pre-market numbers are also flattered because a wider opening range puts the opposite edge further away.
- Many combinations were tested (108-135 per measure), so a few 'significant' differences are expected by chance. Real and placebo breakouts come from the same days, and the four crypto markets move together, so they are not four independent confirmations. The intervals treat the samples as independent.
- The share of days with a breakout is about 99-100 % everywhere because the horizon is several hours; it says nothing about the open.
- The US 500 and US 100 were measured on perpetual contracts that track the indices; at the Tokyo and London opens these trade thinly, outside US cash hours.
- Horizons differ by open (Tokyo to London about 7-8 hours, London to New York about 6.5 hours, New York 6 hours). The placebo uses the same length, so each comparison is fair, but absolute rates are not comparable across opens.
- Weekdays only; crypto weekends were not part of this comparison.
- Earlier and held-out periods mostly agree (100 of 108 pre-market cells, 110 of 135 range cells), but XRP New York moved from 53 % to 63 % (1x pre-market, 15-minute range), and short-series Tokyo cells for silver and oil were unstable.
Cite this
Swiftedge Research (2026). Opening range breakouts: do they follow through? A placebo test. https://swiftedge.uk/research/opening-range-breakouts
<a href="https://swiftedge.uk/research/opening-range-breakouts">Opening range breakouts: do they follow through? A placebo test - Swiftedge Research</a>
Common questions
What is an opening range breakout?
The opening range is the high and low of the first minutes after a session opens. A breakout is when price closes outside it, and many traders follow the direction.
What is the placebo test?
The same breakout rule, started at a random time on the same day. If real opening breakouts do no better than the placebo, the opening adds nothing to the breakout.
Does the opening range breakout work?
Measured in units of its own range, an opening-range breakout was no more likely to keep running than the same breakout at a random time of day: after the New York open, crypto reached one more 15-minute range width 62-67 % of the time, against 75-77 % for the placebo. The table has the numbers for each market and open.
Why does the pre-market hour look better?
The hour before an open is usually quiet, so the open moves more than that hour by itself. That makes multiples of the pre-market range easier to reach; it does not mean the breakout direction predicts anything.