Bitcoin Volatility by Hour: When Does BTC Actually Move?

Bitcoin volatility by hour has a clear shape. Across the last 365 full days (25 Oct 2025 to 24 Oct 2026, UTC), the average one-hour BTCUSDT candle had a range of 0.62%, the busiest hour was 14:00 UTC at 1.01%, and the quietest was 10:00 UTC at 0.46%. The four hours from 13:00 to 16:59 UTC hold about 24% of all hourly range in a day, and weekends are about 40% quieter than weekdays.

This is a data study. We took every hourly candle for one year (8,760 of them, none missing), measured how far each one travelled from low to high, and averaged by hour of day, by weekday, by quarter and by US clock regime. It covers when Bitcoin moves, not where. For a practical guide to the sessions behind these hours, read best time to trade crypto and use the trading sessions tool.

Bitcoin volatility by hour of day

We define the range of a candle as high minus low, divided by its open, in percent. For each of the 24 UTC hours we averaged that number over 365 days. The 14:00 UTC hour means the candle from 14:00:00 to 14:59:59 UTC.

What stands out:

• One clear peak. 14:00 UTC averages 1.01%, followed by 15:00 at 0.96% and 13:00 at 0.81%. Only five hours average above 0.70%: 13:00 to 17:59 UTC.
• The quiet block is the morning. 10:00 UTC is the quietest hour at 0.46%, with 11:00 at 0.49%, 06:00 at 0.50% and 04:00 at 0.50%.
• The busiest hour is 2.2 times the quietest. 1.01% against 0.46%. Outside the 13:00 to 17:59 block, the other 19 hours sit in a narrow band between 0.46% and 0.67%.
• The US session explains the timing, not the size. The NYSE core session opens at 9:30 a.m. ET (NYSE hours and calendars, checked 25 Oct 2026), which is 13:30 UTC on US summer time and 14:30 UTC on winter time. The peak sits right after it. We show the timing match and make no claim about the cause.

Weekday by hour: where the busiest cell is

The same average, split by weekday (UTC dates), gives 168 cells. Brighter means a wider range.

The busiest cell is Tuesday 14:00 UTC at 1.32%. The quietest are Saturday 10:00 and 11:00 UTC at 0.29% each, so the widest cell is about 4.6 times the narrowest. Monday to Friday look alike: the busiest hour is 14:00 UTC on Monday, Tuesday, Friday and Saturday, and 15:00 on Wednesday and Thursday. The one different day is Sunday, whose busiest hour is 22:00 UTC at 0.99%, a Sunday-evening pick-up before the new week.

Weekend vs weekday volatility

Averaged over every hour of the day, the weekday figure is 0.70% and the weekend figure is 0.42%. Weekdays are 1.67 times as active.

Saturday is the quietest day at 0.37%, and Sunday is 0.47%. The five weekdays are within 0.04 percentage points of each other (0.68% on Tuesday to 0.72% on Friday). This is a different measure from the daily-range comparison in weekend crypto trading, which used daily candles over 90 days, so the ratios are not meant to match.

How it looks on a real chart

Here is the most recent full Monday to Friday week, 19 to 23 Oct 2026, in one-hour BTCUSDT candles (Binance spot, fetched 25 Oct 2026). The shaded columns are 13:00 to 17:00 UTC.

1. Tuesday 20 Oct, 14:00 UTC. The widest candle of the week: 1.63% from low to high.
2. Thursday 22 Oct, 01:00 UTC. An exception. The widest hour of that day was in the middle of the night, at 1.16%.

On four of the five days the widest one-hour candle was the 14:00 UTC one (Monday 0.96%, Tuesday 1.63%, Wednesday 1.33%, Friday 1.17%). Across the five days the 13:00 to 16:59 window averaged 0.82% per candle against 0.42% for the other 100 hours. One week is an example, not proof, and Thursday shows that any single day can break the pattern. The year of data is what shows it is a tendency.

Does the busiest hour hold in every quarter?

We split the year into four consecutive blocks of 91 or 92 days and found the busiest and quietest hour in each.

The answer is "mostly, with one shift". The top hour is 14:00 UTC in three quarters and 15:00 UTC in the first. In every quarter, both 14:00 and 15:00 are among the three busiest hours, and the top three always sit between 13:00 and 17:59 UTC. The shift to 15:00 comes in the first window (25 Oct 2025 to 23 Jan 2026), when 15:00 averaged 1.12% against 1.06% for 14:00. That window is almost entirely US winter time, which the next section covers. The second window (24 Jan to 24 Apr 2026) is split between winter and summer time, and 14:00 edged 15:00 there, 1.25% against 1.22%.

The size of the pattern changed far more than its timing. Average hourly range was 0.64%, 0.80%, 0.57% and 0.48% in the four windows, a 1.7-fold swing between the most and least volatile. The quietest hour is the less stable finding: 10:00 UTC was lowest in three quarters, but in the last one 20:00 UTC was lowest at 0.36% and 10:00 UTC was almost level at 0.38%. Treat the quiet hours as a low plateau, not one fixed hour.

Daylight saving: the busiest hour moves with the US clocks

Because the US changes its clocks and UTC does not, the New York open moves by an hour in UTC. We checked whether Bitcoin's hour pattern moves with it. US daylight time ran from 8 March to 1 November 2026 and ended on 2 November in 2025 (NIST, checked 25 Oct 2026). The UK changes its clocks on different dates, 26 Oct 2025 and 29 Mar 2026 (GOV.UK, checked 25 Oct 2026), so to isolate the US effect we left out the 31 days when US and UK clocks disagreed. That leaves two clean periods:

• Summer time: 30 Mar to 24 Oct 2026, 209 days.
• Winter time: 3 Nov 2025 to 7 Mar 2026, 125 days.

Winter was a more volatile stretch overall (average hourly range 0.75% against 0.54%), so the chart divides each hour by the average hour of its own period. 1.00 means an average hour.

The step up moves one hour later in winter. In summer time the 13:00 UTC hour already runs at 1.52 times an average hour and the peak is 14:00 UTC at 0.88% (1.63 times). In winter time 13:00 UTC is an ordinary hour at 1.01 times, 14:00 is 1.64 times and the peak is 15:00 UTC at 1.30% (1.73 times). That is the NYSE opening shifting from 13:30 to 14:30 UTC, and the pattern follows it. It is two periods of one year, not a controlled experiment, so read it as consistent with the US session driving the timing.

In practice: if your own clock or your alerts are set in UTC, the busy block starts at 13:00 UTC until the US changes its clocks on 1 Nov 2026, and at 14:00 UTC after. See session times in winter for the updated London and New York hours and kill zones vs sessions for how this maps to named windows.

Ethereum shows the same shape

We ran the same calculation for ETHUSDT over the same 365 days.

ETH is wider in every hour (0.85% average against 0.62%), but the timing is the same: busiest at 14:00 UTC, quietest at 10:00 UTC, with the weekend about a third quieter (weekday over weekend 1.52 times). By quarter, its busiest hour is 15:00, 14:00, 14:00, 14:00 UTC, matching Bitcoin. In summer time ETH peaks at 14:00 UTC (1.13%) and in winter time at 15:00 UTC (1.81%). Prices and ranges for the coin are on its market page.

The full tables

Average one-hour range, % of open, by UTC hour, 25 Oct 2025 to 24 Oct 2026:

Average one-hour range by weekday (UTC dates), same window:

Method and limits

• Data. Binance spot BTCUSDT and ETHUSDT one-hour klines (Binance Spot API), fetched 25 Oct 2026. The window is the 365 full UTC days from 25 Oct 2025 00:00 to 24 Oct 2026 23:59, which is 8,760 candles per pair with no gaps.
• Range. (high − low) ÷ open of each candle, in percent, averaged as a plain mean. A few violent hours raise a mean, so the figures describe typical behaviour including the spikes.
• Hour and day. The hour is the candle's open time in UTC. A weekday is the UTC date. Moscow time (MSK) is UTC+3 all year, so 14:00 UTC is 17:00 MSK.
• Quarters. Four consecutive blocks of 91, 91, 91 and 92 days, not calendar quarters.
• One venue, one year. Binance spot only. Perpetual futures, other exchanges and other years can give different numbers, and one year is a single regime. Volatility swung 1.7-fold between our quarters.
• Range is not direction. A wide candle can be up or down. Nothing here says which way price will go or that a busy hour is a good time to trade. It also does not measure volume, spreads or liquidity. Volatility is also a quarter of the weight in the Fear and Greed Index, which is why that index tends to repeat what the chart shows.
• No cause is proven. The match between the peak and the US stock-market open is a timing observation.

Using it

Treat the table as a map of when moves are likely to be larger, so you can size positions and place stops for the hour you are in. The position size calculator turns a stop distance into a size, and the Bitcoin market page shows the current structure. More studies are in the Learn ICT hub.

FAQ

What time is Bitcoin most volatile? In the 365 days to 24 Oct 2026, the busiest hour for BTCUSDT was 14:00 UTC, with an average one-hour range of 1.01%. 15:00 UTC was next at 0.96%. The busiest single weekday-hour cell was Tuesday 14:00 UTC at 1.32%.

What time is Bitcoin least volatile? 10:00 UTC, at an average range of 0.46%. Saturday 10:00 and 11:00 UTC were the quietest cells of the week at 0.29%. In the latest quarter 20:00 UTC was marginally lower, so the quiet hours are a plateau rather than one fixed time.

Is Bitcoin more volatile on weekends? No. Over the same year, the average hourly range was 0.42% at weekends against 0.70% on weekdays. Saturday was the quietest day at 0.37%. The exception is Sunday evening, where 22:00 UTC averaged 0.99%.

Does the busiest hour change with daylight saving time? Yes, by one hour. In the 209 clean days of US summer time the peak was 14:00 UTC, and in the 125 clean days of winter time it was 15:00 UTC. The US clock change on 1 Nov 2026 moves the New York open from 13:30 to 14:30 UTC.

Is the pattern the same for Ethereum? Yes. ETHUSDT was busiest at 14:00 UTC (1.35%) and quietest at 10:00 UTC (0.64%), with the same one-hour shift between summer and winter time.