---
title: Does Bitcoin always fill CME gaps? A spot-price test
url: https://comparand.xyz/blog/does-bitcoin-fill-cme-gaps
description: Coinbase Bitcoin returned to its Friday price within 3 days after 71.81% of weekends, but a same-distance level was touched about as often.
updated: '2026-10-21'
category: Patterns
---

# Does Bitcoin always fill CME gaps? A spot-price test

Coinbase Bitcoin returned to its Friday price within 3 days after 71.81% of weekends, but a same-distance level was touched about as often.

No, not always, and not more often than chance. In our test, Coinbase Bitcoin spot price came back to its Friday level within 3 days after 71.81% of weekends (415 weekends with a complete 3-day window, December 2017 to 2025), and within 14 days after 86.73%. But a level the same distance away was touched about as often on weekends with similar volatility. So we found no special "gap fill" effect.

One limit first: this test uses spot prices, not CME futures prices. It answers "how often did spot return to its Friday price?", not "how often did a CME futures chart fill?"

## What is a CME gap, and what did we measure?

CME is a big US exchange that trades Bitcoin futures. A futures contract is an agreement linked to a coin's price. CME's Bitcoin futures stop trading over the weekend. We used the regular break from Friday 4:00 pm to Sunday 5:00 pm Chicago time.

Bitcoin itself keeps trading all weekend, usually more calmly (see [Is crypto more volatile on weekends?](https://comparand.xyz/blog/crypto-weekend-volatility)). So a CME futures chart can show an empty space between Friday's last price and Sunday's first price. People call that space a gap, and many say it always gets filled.

We cannot see CME's own prices. So we measured something close but different: how far Coinbase spot moved while CME was closed, and whether spot touched its Friday level again. "Spot" means the coin itself, priced in US dollars. The Friday level is the price at the end of the hour when CME closed.

*Diagram: One real weekend on a timeline from Friday 4:00 pm to Sunday 5:00 pm Chicago time: Coinbase spot went from $84,192.75 on 2025-02-28 to $94,230.48 at the reopen, a +11.92% move while CME was shut.*
## How often did spot come back to its Friday price?

Here are all Coinbase BTC-USD weekends from the first Friday after CME Bitcoin futures began trading (December 2017) through 2025. A weekend counts as "back" if any later hourly high or low reached the Friday price within the window. A single wick counts.

| Time after the Sunday reopen | Weekends scored | Back at the Friday price |
| --- | --- | --- |
| 1 day | 418 | 55.98% |
| 3 days | 415 | 71.81% |
| 7 days | 405 | 80.25% |
| 14 days | 392 | 86.73% |

Weekends with a missing hour were dropped for that window, so the counts shrink as the window grows. Among weekends that came back within 14 days, the middle wait was 12 hours.

A check on Binance's BTCUSDT perp (a futures contract that never expires) over 329 weekends from September 2019 to 2025 gave 59.27% at 1 day, 74.77% at 3 days, 81.16% at 7 days and 86.93% at 14 days. Close to Coinbase, but a different market, so we do not add them together.

## Does the size of the weekend move matter?

Yes. A small move is easy to undo. A big one is not. From here on, figures cover all 461 weekends from December 2017 to October 2026. That includes 42 weekends in 2026, where we assumed the regular CME break still applied (we did not check). The middle weekend move (ignoring direction) was 1.16%. One weekend in ten moved more than 4.74%. Spot was higher at the reopen on 59.44% of weekends.

| Move while CME was closed | Weekends | Back within 1 day | Back within 3 days | Back within 14 days |
| --- | --- | --- | --- | --- |
| Under 1% | 211 | 78.20% | 90.00% | 95.03% |
| 1% to 3% | 147 | 46.26% | 70.14% | 89.05% |
| Over 3% | 103 | 26.47% | 41.18% | 66.30% |

Counts can be a few weekends lower where bars were missing. After a move over 3%, about 1 in 3 weekends had not come back after 14 days. When price did come back within 14 days, the middle wait was 2 hours after moves under 1%, 19 hours after moves of 1% to 3%, and 45 hours after moves over 3%.

## Is that more than chance?

This is the question that matters. Price that moves a lot touches many levels, whether or not one of them is a "gap".

So we ran a placebo test. For each real weekend, we took the same size of move, in the same direction, and placed it against the Sunday reopen price of a different weekend. Then we asked how often price touched that level in the same time.

Against random weekends, big moves look special. That is a trap: big moves tend to happen on volatile weekends, and volatile weekends touch more levels anyway. So we compared only with weekends that had similar volatility before the Friday close.

| Move while CME was closed | Back within 3 days (real) | Back within 3 days (similar-volatility weekends) |
| --- | --- | --- |
| Under 1% | 90.00% | 89.63% |
| 1% to 3% | 70.14% | 70.55% |
| Over 3% | 41.18% | About 40% |

The rows are almost the same. For moves over 3%, the real share was about 1 point higher; with about 102 weekends, the usual wobble is about 5 points. On Binance it was about 5 points higher, with a wobble of about 6. Both are inside the noise. With 12 size-and-window cells per market, some will look strong by luck (see [how many trades a backtest needs](https://comparand.xyz/blog/how-many-trades-backtest-sample-size)).

One more check points the same way. Before CME Bitcoin futures existed, there was no CME gap at all. Yet Coinbase spot still came back to its Friday 4:00 pm price within 3 days after 54.92% of 126 weekends. So price can return to an old level with no CME gap at all.

## What happened after the ten biggest weekend moves?

These are the ten largest spot moves while CME was closed. They are the cases people remember.

| Friday | Move | Friday price to reopen price | Back within 14 days |
| --- | --- | --- | --- |
| 2019-05-10 | +13.69% | $6,293.31 to $7,155.00 | No |
| 2019-05-17 | +13.64% | $7,107.19 to $8,076.41 | No |
| 2021-01-01 | +13.37% | $29,253.63 to $33,166.13 | Yes |
| 2020-05-08 | −12.74% | $9,999.99 to $8,726.08 | No |
| 2019-10-25 | +12.40% | $8,568.00 to $9,630.05 | Not scored |
| 2025-02-28 | +11.92% | $84,192.75 to $94,230.48 | Yes |
| 2020-03-27 | −11.77% | $6,682.00 to $5,895.61 | Yes |
| 2018-06-08 | −11.61% | $7,646.82 to $6,758.78 | No |
| 2019-07-12 | −10.55% | $11,915.42 to $10,657.96 | No |
| 2020-03-06 | −9.60% | $9,137.51 to $8,260.00 | No |

Of the 9 we could score, 3 came back within 14 days. A big gap can stay open for a long time.

*Diagram: Show the ten largest weekend moves as paired Friday and reopen prices, marking which came back to the Friday level within 14 days.*
## Did the pattern change by year, and what about Monday?

From 2018 to 2025, the 3-day share ranged from 61.54% in 2019 to 78.85% in 2024. Each year has about 50 weekends, so those year-to-year swings are partly luck.

We also looked at Monday. After a weekend where spot rose, the next UTC Monday fell 52.19% of the time (274 weekends), against 50.76% for all 461 Mondays. That is no real difference. For more on Mondays, see [Bitcoin Sunday pump, Monday dump](https://comparand.xyz/blog/sunday-pump-monday-dump).

Last, we tested a simple rule on past weekends: at the Sunday reopen, move toward the Friday level, and exit 72 hours later. Over 460 trades, the average trade lost 0.07% before fees and 0.17% after an assumed 0.05% fee on each side. This describes old weekends, not a plan to trade.

## What does this not show?

- It does not show what CME futures prices did. We did not use them.
- The share before CME futures (54.92%) was lower than after (71.81%). We cannot say why; the market was very different then.
- Weekends are not fully separate. Nearby weekends can share one trend.
- Past cases, not a forecast.

## How we measured this

We used hourly Coinbase BTC-USD spot bars and a Binance BTCUSDT check. For each weekend, we took the close of the hour ending at Friday 4:00 pm Chicago time as the Friday price, and the close of the hour ending at Sunday 5:00 pm Chicago time as the reopen price. A weekend is "back" when a later hourly low (for an up move) or high (for a down move) reaches the Friday price. We treated 13 CME holiday Fridays as normal; removing them changed the results very little. Hourly bars can hide a short wick, and we did not use CME prices.

## Try this in Comparand

Ask your AI agent: "Use venue_markets to find BTC-USD on Coinbase, then venue_bars for hourly bars on my dates."

- **venue_markets** lists the markets one venue offers, with a filter by kind or search word. It is free.
- **venue_bars** returns the open, high, low, close and volume for each hourly bar, completed bars only. It is free too.
- Your agent finds the Friday and Sunday bars, works out the move, and checks later highs and lows. The tool returns bars, not a gap report.

Each answer says which data it came from and what the call cost. You can also [try 3 coins free](https://comparand.xyz/try) to see past charts and what followed them, and the [docs](https://comparand.xyz/docs) show how to connect your agent.

## The short version

- Coinbase Bitcoin spot came back to its Friday level within 3 days after 71.81% of weekends, and within 14 days after 86.73%.
- Bigger moves came back less often, but a same-distance level on a similar-volatility weekend was touched about as often.
- The ten biggest weekend moves are mixed: 3 of the 9 we could score came back within 14 days.
- Moving toward the Friday level lost money on average in past weekends.
- This is a spot-price test, not CME futures prices.

## Questions people ask

**Does Bitcoin always fill CME gaps?**

No. In our spot-price test, Bitcoin came back to its Friday level within 14 days after 86.73% of weekends, so about 1 in 8 did not. The test used Coinbase spot prices, not CME futures prices, so it cannot say what a CME futures chart did.

**How often does Bitcoin return to its Friday price within a week?**

On Coinbase BTC-USD spot, from December 2017 through 2025, price touched its Friday level again within 7 days after 80.25% of 405 weekends with a complete week of prices. These are past cases, not a forecast.

**Are bigger weekend moves filled less often?**

Yes, and that is mostly what you would expect from distance alone. After moves over 3%, price was back at its Friday level within 3 days after 41.18% of 102 weekends. After moves under 1%, it was 90.00% of 210. A same-distance level on weekends with similar volatility was touched about as often.

**Did trading toward the Friday price make money in your test?**

No. We tested 460 weekends where we moved toward the Friday level at the Sunday reopen and exited 72 hours later. The average trade lost 0.07% before fees and 0.17% after an assumed 0.05% fee on each side. This describes the past, not a tradable plan.