Do weekend gaps fill?
Yes. About 89% of them close back to Friday's price within a day. That is the headline of the most-sold report in this category, and it is true. It is also empty: an ordinary midweek move of the same size fills just as often. Here is that second number, which nobody prints — from 11,773 weekends across 12 FX and metals markets.
The claim
Search "weekend gap" and the same page comes back in a dozen liveries: forex gaps fill 90% of the time, so fade the gap on the Sunday reopen. The number is usually right. The reasoning has a hole in it that a single control closes.
What we measured it against
For every weekend gap we found a size-matched midweek move: an hour whose next hour travelled 0.85–1.35× as far as that gap did. The reference is taken from the bar after the move, so the level we ask about is never contained in the move by construction. Then we asked the same question of both — did price return to the starting level within 4, 12 and 24 hours? Same markets, same years, same definition of "filled". Only the cause of the move differs.
The result: the gap is not special
| Horizon | Weekend gaps filled | Matched midweek moves | Difference |
|---|---|---|---|
| Within 4 hours | 75% | 70% | +5 pts |
| Within 24 hours | 89% | 88% | +1 pts |
At a day out — the horizon the famous number quotes — there is nothing there: 89% against 88%. The one honest finding is at the short end, where gaps fill about 5 points more often than a matched move over the first four hours. That is real, it is small, and the next section is about whether you can be paid for it.
⚠ And it is not uniform. 10 of the 12 markets fill more often than their own control at four hours; 2 fill less often — EUR/USD among them, at 69.2% against 69.9%. A pooled figure is an average of markets that disagree, which is why the table below is the part worth reading.
Every market, with its own control
| Market | Weekends | Median gap | Gap filled ≤4h | Matched move | Δ | Gap filled ≤24h | Matched move | Δ | Clears reopen spread |
|---|---|---|---|---|---|---|---|---|---|
| EUR/USD | 1,187 | 5.8 pips | 69.2% | 69.9% | -0.7 | 87.3% | 88.8% | -1.5 | 52.5% |
| GBP/USD | 1,041 | 8.5 pips | 71.3% | 69.9% | +1.4 | 88.5% | 88.6% | -0.1 | 65.9% |
| USD/CHF | 643 | 6.5 pips | 77.0% | 67.6% | +9.4 | 91.0% | 88.4% | +2.6 | 55.5% |
| USD/JPY | 915 | 7.6 pips | 72.1% | 68.7% | +3.4 | 85.4% | 88.0% | -2.6 | 54.9% |
| AUD/USD | 1,157 | 7.4 pips | 68.5% | 70.0% | -1.5 | 84.4% | 87.1% | -2.7 | 70.4% |
| USD/CAD | 1,190 | 6.5 pips | 72.9% | 70.2% | +2.7 | 90.6% | 90.3% | +0.3 | 55.0% |
| NZD/USD | 1,039 | 6.4 pips | 75.4% | 69.2% | +6.2 | 87.5% | 87.2% | +0.3 | 50.1% |
| EUR/JPY | 1,049 | 10.9 pips | 72.4% | 68.8% | +3.6 | 85.1% | 87.3% | -2.2 | 60.0% |
| GBP/JPY | 829 | 14.8 pips | 75.0% | 68.3% | +6.7 | 87.1% | 87.7% | -0.6 | 68.9% |
| EUR/GBP | 491 | 4.7 pips | 76.0% | 72.6% | +3.4 | 90.6% | 87.8% | +2.8 | 45.8% |
| Gold (XAU/USD) | 1,201 | 6.86 pips | 81.5% | 77.9% | +3.6 | 90.8% | 91.4% | -0.6 | 41.5% |
| Silver (XAG/USD) | 1,031 | 3.7 pips | 84.5% | 81.6% | +2.9 | 92.5% | 93.4% | -0.9 | 20.2% |
Δ is the gap's fill rate minus the matched move's. Positive means the gap filled more often than an ordinary move of the same size.
The indices do the opposite, and that is the interesting part
Stock indices gap at every session boundary, not only over a weekend, so we counted them separately. They invert the folklore: an index gap fills less often than a matched ordinary move — 55% against 69% at four hours. An index that gaps tends to keep going, which is the reverse of what fade-the-gap advice tells you to do.
| Market | Weekends | Median gap | Gap filled ≤4h | Matched move | Δ | Gap filled ≤24h | Matched move | Δ | Clears reopen spread |
|---|---|---|---|---|---|---|---|---|---|
| US30 (Dow) | 573 | 29.02 pts | 51.1% | 65.4% | -14.3 | 80.3% | 85.8% | -5.5 | 88.0% |
| NAS100 | 343 | 22.88 pts | 56.6% | 69.1% | -12.5 | 82.8% | 85.4% | -2.6 | 89.2% |
| SPX500 | 343 | 5.3 pts | 54.5% | 68.5% | -14.0 | 82.5% | 86.7% | -4.2 | 77.0% |
| GER40 (DAX) | 527 | 21.95 pts | 52.2% | 66.2% | -14.0 | 79.1% | 83.1% | -4.0 | 89.8% |
| UK100 (FTSE) | 191 | 6.89 pts | 62.3% | 72.8% | -10.5 | 89.5% | 89.1% | +0.4 | 63.4% |
And then there is the cost
The Sunday reopen is the widest spread of the week, and it is widest in its first hour. We charge each gap the round trip of the reopen hour it landed in — measured from bid and ask quotes in the 10 currency pairs we hold them for, an assumed 3× the weekday figure in the metals, which the archive does not cover. Against that, only 55% of weekend gaps are even large enough to be worth trying to fade: the median FX gap is about 6.86 pips. A strategy built on the strongest version of this finding is playing for a few points, on the fraction of weekends that clear the toll, in the hour of the week when the toll is highest.
What we would actually say
- "Gaps fill 90% of the time" is true and tells you nothing. So does everything else that moves.
- There is a small real effect at the short end in FX — a few points over the first four hours — and it is smaller than the reopen spread for most gaps.
- Do not carry the rule into indices. There the sign flips.
- Crypto is excluded entirely — it has no weekend, which is a different question and one we are measuring separately.
The method, the null and the raw counts are in section 5 of the evidence page. The data behind it is free to download and redo.
Other reports this category sells, tested
Every claim we have tested, including the ones that held, is on the reports index.
We publish the tests that failed as loudly as the ones that held. Every number on this site carries its sample size, and every claim carries what it was measured against.
Read the whole evidence page →