Do Bitcoin CME gaps always fill? 440 weekends tested against chance (2017–2026)

What a CME gap is

Bitcoin trades every minute of every day on crypto exchanges. Until May 2026, the Bitcoin futures on the Chicago Mercantile Exchange (CME) did not. They stopped on Friday afternoon, New York time, and reopened on Sunday evening. Whatever Bitcoin did over the weekend showed up on the CME chart as a jump between Friday's last price and Sunday's first price: a CME gap.

Few ideas are repeated more often in crypto trading than "CME gaps always fill": sooner or later, the price comes back to Friday's close. Traders post charts of open gaps as targets, sometimes years old.

On May 29, 2026, CME moved its crypto futures to 24/7 trading, and the weekend gap mostly disappeared. That gives us a closed sample: 440 weekends, from the first CME Bitcoin futures in December 2017 to May 22, 2026. We measured every one, and asked two questions. How often, and how fast, did the price come back to Friday's close? And is that any different from what Bitcoin does on any other day?

How we measured the gaps

We do not have CME's own tick data, so we used the price everyone actually trades against: Binance BTC/USDT hourly candles. For each weekend we took the Bitcoin price at the moment CME closed on Friday (5 pm New York time, which is 21:00 or 22:00 UTC depending on US daylight saving) and the price when CME reopened on Sunday (6 pm New York). The difference is the gap. The gap is "filled" when a later hourly candle trades back through Friday's price.

CME futures trade at a small premium or discount to spot, so their own gaps are slightly different from ours. But the question traders ask is about Bitcoin's price coming back to a level, and that is what this measures.

Most gaps filled, and most of them quickly

The headline claim is roughly true. Of the 440 weekend gaps, 96.8% were filled at some point. 55.5% were filled within 24 hours of the Sunday reopen, 79.5% within a week, and 88.6% within 30 days. 17.5% were filled in the very first hour, before most traders had even looked at the chart.

But size matters a lot. Gaps under 0.5% (28% of all weekends) were filled within a day 85% of the time, with a median fill time of one hour. Gaps above 5% (about one weekend in ten) were filled within a day only 17% of the time, and the median wait was 370 hours, over two weeks. Some took much longer: the gap from December 3, 2021, with Bitcoin at $53,610, took 813 days to fill, and one from January 2018 took 1,044.

The test that matters: is this special to gaps?

Here is the problem with "gaps always fill". Bitcoin moves around constantly. Any price it traded at a day ago, it has a good chance of trading at again soon. So a high fill rate on its own proves nothing. What we need is a comparison.

We did it two ways. First, a random baseline: for each gap, we picked 20 random weekday hours and asked how often Bitcoin touched a level the same distance away, in a random direction. Second, a midweek placebo: we measured the same 49-hour window from Tuesday evening to Thursday evening every week, as if CME closed in the middle of the week, and treated that move as a "gap".

For small gaps, there is nothing special at all. Gaps under 1% filled at almost exactly the same rate as random levels the same distance away (65% vs 77% within a day for 0.5–1%) and as the midweek placebo (69%). If anything, weekend gaps of this size filled a little less often than chance.

For bigger gaps, there is a real difference. Weekend gaps of 2–5% were filled within 24 hours 41% of the time, against 22% for midweek moves of the same size and 28% for random levels. Within a week the numbers were 73% vs 59% and 61%. For gaps above 5%, 37% filled within a week, against 33% for the placebo and 32% for random levels, so only a small edge.

Why would 2–5% weekend gaps come back more than midweek moves? A likely reason is liquidity. Weekend trading is thinner, so moves on Saturday and Sunday can run further than the news justifies, and when Wall Street and the big desks return on Monday, part of the move is reversed. That is a weekend effect, not a magic property of the CME chart.

Up gaps and down gaps

Gaps up (Bitcoin rose over the weekend) were more common: 257 against 183 down gaps. But they behaved almost the same. Up gaps were filled within a day 54% of the time and down gaps 58%; within a week, 79% and 81%. We also checked whether the gap direction carried on into Monday. For gaps of 1% or more, Bitcoin's first 24 hours after the reopen moved in the same direction as the gap 53% of the time, which is basically a coin flip.

Gaps got smaller over the years

The typical gap shrank as Bitcoin grew up. In 2018 and 2019 the median gap was about 2%. From 2022 onwards it was usually under 1%. Gaps also got filled within a week a bit more often: 76% in 2018–2021, 83% in 2022–2026. Smaller, faster-filling gaps are consistent with a market where weekend trading is less extreme than it used to be.

The gaps that never filled

Fourteen gaps (3.2%) had not been filled by November 5, 2026. Most of them are up gaps from cheaper times: Bitcoin went up over a weekend in 2018–2024 and never came back down to the Friday price. The highest is the 3% gap from September 6, 2024, at about $52,900. For it to fill, Bitcoin would have to fall more than 30% from where it is now. These are the "unfilled CME gaps" you see on charts, and some may never fill.

Notice what that list says about the rule. A gap is only "unfilled" while the price keeps rising away from it. In a market that rose about five times from the first CME weekend to today, and more than 20 times from its 2018 low, it was always more likely that old down gaps would fill (the price comes back up) than that old up gaps would.

What changed in May 2026

CME's 24/7 crypto trading started on May 29, 2026. Weekend trades now go into the following business day's clearing, and CME keeps a short weekly maintenance window, so small pauses still exist, but the large Friday-to-Sunday jump on the CME chart is gone. That means:

• Old open gaps on charts are now historical levels, not "gaps" waiting for a mechanism to close them.
• Any weekend effect that remains should show up in Bitcoin's own weekend price action, not on a CME chart. Our study on stock perps and the Monday open looks at the same weekend question for stocks.

What to take from this

1. The fill rate is mostly what Bitcoin does anyway. For gaps under 1%, a weekend gap was no more likely to fill than any random level at the same distance.
2. Medium gaps (2–5%) did come back more often than chance, about twice as often within a day as a midweek move of the same size. That fits thin weekend trading being partly corrected on Monday.
3. Big gaps are a poor short-term target. Above 5%, the median wait was over two weeks, and a leveraged position aimed at a quick fill could be liquidated first.
4. The classic CME gap is over. Since late May 2026 there is no regular weekend gap on CME Bitcoin futures.

You can follow Bitcoin's current price, trend and levels on our Bitcoin price analysis page. Kumocast provides free market data and analysis, not trading signals or financial advice.