Best time to trade crypto: volatility by hour and time zone

Crypto trades around the clock, but it moves on New York time

Spot crypto markets never close, yet activity is far from even across the day. Across BTC, ETH, SOL, XRP and DOGE on weekdays from January 2024 to September 2026, the average hourly price move peaked in the hour starting 14:00 UTC at 1.71 times its overall average, and the four hours from 13:00 to 17:00 UTC were all at least 1.28 times. Trading volume was concentrated in the same hours: the 14:00 UTC hour traded 1.82 times an average hour of the same day.

Those hours are the US morning. Measured in New York local time, the peak is the hour from 10:00 to 11:00 a.m., at 1.80 times average, which is sharper than the peak measured in UTC because the busy period moves with US daylight saving. The quietest weekday hours run from the US late evening into the Asian working day: 23:00 UTC averaged 0.79 times, and 04:00 to 12:00 UTC stayed between 0.86 and 0.99 times.

The busiest and quietest hours in your time zone

The table converts the busiest and quietest four-hour weekday windows into local time. They fall at 9 a.m. to 1 p.m. in New York, 2 to 6 p.m. in London and Lagos, 6 to 10 p.m. in Pakistan, about 6:30 to 10:30 p.m. in India, 9 p.m. to 1 a.m. in Manila, Kuala Lumpur, Singapore and Hong Kong, and around midnight to 4 a.m. in Sydney.

Where clocks do not change, the busy window moves one hour later when US clocks go back in November and earlier again in March. In India, for example, the US stock market opens at 7:00 p.m. during US summer time and 8:00 p.m. in winter. A second, smaller rise comes at 00:00 to 02:00 UTC (1.16 to 1.20 times), around the daily candle close and the start of the Asian morning.

Weekends are quieter, not closed

Bitcoin’s average weekend daily volume was 49% of a weekday’s, and its average hourly move was 41% smaller. The weekend drop was smaller for the other coins: daily volume was 59% of weekdays for ETH, 67–68% for XRP and SOL and 76% for DOGE. The US-morning peak largely disappears on Saturdays and Sundays; no weekend hour averaged more than 0.99 times the overall level.

Lower activity also means thinner markets, so weekend moves can still be sharp when news arrives. Our CME gap study looks at what happened when weekend moves met the reopening of regulated futures.

No hour of the day reliably goes up

Timing volatility is not the same as timing direction. Bitcoin’s average return in each UTC hour ranged from −2.7 to +3.7 basis points (hundredths of a percent). The hour starting 21:00 UTC averaged +3.7 bp (t 2.44) and the hour starting 23:00 UTC −2.7 bp (t −2.11); both kept their sign in each of 2024, 2025 and 2026.

With 24 hours tested, one or two results of that size are expected by chance, and both are far below the cost of a round trip at typical taker fees of around 0.1% per side. The useful pattern is when markets are active, not which way they move.

Is the pattern stable?

The peak hour was the same in April 2022 to December 2023: 10:00 to 11:00 a.m. New York time. It was less pronounced, at 1.31 times average against 1.80 times since 2024. US trading hours have become more dominant since US spot Bitcoin ETFs began trading in January 2024, although these data cannot show that the ETFs caused the change.

How the numbers were calculated

We used confirmed OKX spot USDT hourly candles for BTC, ETH, SOL, XRP and DOGE from 1 January 2024 to 23 September 2026, keeping only UTC days with all 24 candles (994 days per coin). An hour’s move is its absolute log return, divided by that coin’s average over the whole period; an hour’s volume is base volume times the closing price, divided by the average hour of the same day. Each coin is averaged by hour first, then the five coins are weighted equally.

Local windows convert every candle to local time with the relevant daylight-saving rules and pick the four consecutive weekday hours with the highest and lowest average move. In half-hour time zones such as India, a UTC candle starts at half past the local hour. Spreads, order-book depth and fees were not measured directly.

What is the best time of day to trade crypto?

For the most movement and volume, weekday US morning hours: 9 a.m. to 1 p.m. in New York (13:00–17:00 UTC in US summer, an hour later in winter), roughly 2–6 p.m. in London and Lagos and 6:30–10:30 p.m. in India. No hour had a reliable upward or downward bias after trading costs.

Does crypto trade 24/7, including weekends?

Yes. Spot crypto markets do not close on weekends or holidays. Activity is lower, though: Bitcoin’s weekend daily volume averaged 49% of a weekday’s, and its average hourly price move was 41% smaller.

What time does the crypto daily candle close?

Most exchanges and charts close the daily candle at 00:00 UTC: 8 p.m. in New York (7 p.m. in winter), 1 a.m. in London in summer (midnight in winter), 5:30 a.m. in India and 8 a.m. in Manila, Singapore and Hong Kong.

Limitations

Five large coins on one exchange, spot candles, January 2024 to September 2026. Hourly candles miss timing within the hour, and local windows average US summer and winter time. Spreads, depth and fees were not measured directly. Activity patterns can change and are not trading advice.

Sources and data access

• OKX market data documentation
• NYSE: holidays and trading hours
• OKX trading fee schedule

Cited sources document external facts or data formats; they do not endorse our findings. The public snapshot contains the table and methods. Source-file fingerprints are included for recalculated datasets. Full archived raw files are not distributed here; fingerprints identify the files used, rather than independently proving the results.

Aggregate data snapshot (JSON) · Markdown · Methods and limitations

Live prior 30-day hourly baseline — separate from the historical study

Only the current watch list is shown, up to 12 assets. Missing or stale estimates remain unavailable. Venue and quote currency are kept separate; beta needs matching BTC data. Hourly statistics are not minute-move probabilities.

Observation cutoff (UTC): 24/10/2026, 15:53

Related research