When is the crypto market actually liquid?

Bitcoin volume peaks around 14:00 to 15:00 UTC, but displayed depth is strongest earlier. We measured spread, book depth, volume and volatility by hour.

The direct answer

For Binance BTC perpetuals, the busiest hours were not the deepest hours.

On weekdays, traded volume peaked at 14:00 and 15:00 UTC at 2.72 and 2.61 times the day's median. Realized volatility was also highest around that window. Displayed book depth within 0.1% of price was slightly below its daily median.

The deepest recurring BTC hour was 11:00 UTC at 1.066 times the day's median depth. The broader 04:00 to 12:00 window was generally deeper and calmer even though it traded less volume.

The clearest thin hour was 21:00 UTC, when BTC depth fell to 0.923 times the daily median. At 22:00 it remained below the median at 0.936, but 21:00 is the stronger result rather than a universal two-hour window.

Quoted spread barely changed by hour because BTC perpetuals usually sat at the minimum price tick. Volume said when the market was active. Book depth said when more displayed size was available near price. They did not give the same answer.

Liquidity is not one number

Traders often use liquidity to mean several different things:

• Tight spread: the best bid and ask are close together.
• Book depth: more resting size is available near the current price.
• Trading activity: many trades and high volume are going through.
• Low price impact: an order can execute without moving price much.
• Fast recovery: the book refills after size trades.

These properties can move together, but they do not have to.

A market can trade enormous volume because news has arrived and price is moving quickly. That is an active market, not automatically an easy one. Resting liquidity can pull back while aggressive orders race through the book.

A quieter market can display more depth and lower volatility. That may be friendlier for a moderate order, but the displayed size can still cancel before execution and the low activity can make a larger position harder to exit.

The question "When is crypto liquid?" therefore needs several measurements, not a volume chart alone.

What spread can and cannot show

Quoted spread is the distance between the best bid and best ask. In basis points:

spread bps = (best ask - best bid) / midpoint x 10,000

Narrower is usually better for an immediately executable small order.

In a very liquid market, the spread often sits at one exchange tick. Once it reaches that floor, it cannot narrow further even if competition and available size improve. Hourly spread then looks flat while depth and impact continue to change.

That is exactly what happened in the BTC perpetual sample. The median spread was about 0.011 basis points, and the normalized hourly medians stayed within roughly 0.1% of the day's median. Spread alone could not distinguish the busy US window from the deeper pre-US book.

What depth adds

We measured the combined displayed bid and ask size within 0.1% of the midpoint, converted into dollars at the contemporaneous price.

This answers a more useful question than top-of-book size alone:

> How much displayed liquidity is available near the market before price moves > roughly 10 basis points away from the midpoint?

Deeper is not a guarantee of execution. Orders can cancel, hidden liquidity is not visible, and a fast market can consume the band before a trader reacts. But near-price depth separates hours that share the same one-tick spread while offering very different displayed capacity.

Why time of day still matters in a 24-hour market

Crypto never closes, but the participants around it follow human and institutional schedules.

Regional business hours overlap. Traditional markets open and close. Economic data arrives on a calendar. Dealers adjust risk around settlement and staffing changes. Weekday participation differs from weekend participation.

Those forces create a repeating intraday pattern without creating a universal "best session." The same hour can be deep on a normal day and disorderly during a liquidation or macro release.

UTC also stays fixed while local daylight-saving schedules move. A US cash market event can shift by one UTC hour across the year. Hourly averages blur that seasonal change.

What the weekday shape looks like

The BTC pattern has four broad parts.

00:00 to 02:00 UTC: active but not deep

Volume sits near the daily median and volatility is slightly elevated. Depth is about 3% to 6% below the day's median.

04:00 to 12:00 UTC: deeper and calmer

Displayed depth is usually 2% to 7% above the daily median. Volume and realized volatility are lower. The single deepest median hour is 11:00 UTC.

13:00 to 17:00 UTC: the activity surge

Volume rises sharply and peaks at 14:00 to 15:00. Volatility rises with it. Depth falls back toward or below the daily median.

This is the clearest example of activity and liquidity separating. More trades occur, but the nearby displayed book is not at its deepest.

21:00 UTC: the clearest thin hour

Depth is weakest at 21:00, and volume is also low. The historical sample keeps depth below its daily median at 22:00, but the 21:00 result is the more robust hour to carry forward. At 23:00 volatility falls on weekdays, but the market still does not display the deeper book seen earlier in the day.

A practical way to use the clock

This study describes average market conditions. It does not test a trading strategy or prove that a specific order receives better execution at a fixed hour. Use the pattern as a baseline that current conditions can confirm or override.

1. Decide which liquidity question you have. A small market order cares about spread and top levels. A larger order cares about depth, impact and recovery. A signal trader may care more about activity and volatility.

2. Compare the current hour with its normal state. If 14:00 UTC is usually busy but today's volume is dead, the calendar pattern is not the current market. Live conditions take priority over the average.

3. Treat high volume and high volatility together. A burst of activity can offer fills while increasing impact and adverse selection. Volume alone does not tell you the cost.

4. Check both sides of the book. Total depth can hide an imbalance. A deep bid and thin ask behave differently from balanced depth even if the combined dollars match.

5. Separate weekdays from weekends. The weekday US-hour activity peak is much larger. Weekend depth and volatility follow a different shape, especially late in the UTC day.

6. Size from the live book, not the historical average. The table below is context. Execution decisions need current spread, depth and recent refill behavior on the venue being traded.

How we measured the day

The primary sample uses Binance BTC perpetual one-minute order-book depth and trade candles from January 28, 2024 through March 11, 2026.

For every UTC hour we measured:

• median quoted spread in basis points;
• median combined bid and ask depth within 0.1% of mid, converted to dollars;
• traded dollar volume;
• realized volatility from one-minute returns.

Market size and price changed substantially during the sample. Comparing raw dollars by hour would let the larger recent era dominate. Each hourly value was therefore divided by that same UTC day's median before hours were combined.

A normalized depth of 1.10 means the hour displayed 10% more near-price depth than the median hour of its own day. A normalized volume of 2.00 means twice that day's median hourly volume.

Only days with all 24 hourly observations were included. The BTC sample has 287 complete weekdays and 148 complete weekend days. ETH contributes 411 complete days and SOL 654 as robustness markets. Uncertainty was estimated by resampling complete UTC days.

BTC weekday results

The depth differences are not only ordering noise. At 11:00 UTC the day-block 95% interval for normalized depth is 1.055 to 1.075. At 14:00 it is 0.966 to 0.984, and at 21:00 it is 0.907 to 0.942.

ETH and SOL repeat the broad shape. Their weekday books are deepest around 09:00 to 12:00 UTC, volume and volatility peak around 14:00 to 15:00, and depth falls sharply late in the UTC day, with 21:00 among the weakest hours.

Weekends are flatter during the US window

BTC weekend volume at 14:00 UTC is 1.49 times the weekend daily median, far below the 2.72 weekday multiple. Depth is near its daily median rather than falling as sharply.

The late-day weakness remains. BTC depth at 21:00 and 22:00 UTC is 0.920 and 0.911 times the weekend median. Weekend volatility rises into 22:00 and 23:00 even though volume is not at a weekday-style peak.

The result is not "weekends are always illiquid." It is that the strong weekday activity cycle weakens, while the late-UTC thin-book window remains visible.

Common mistakes

• Calling volume liquidity. Volume measures completed activity, not the cost of the next trade.
• Using spread alone in a one-tick market. Once spread reaches its floor, depth and impact can still vary.
• Treating displayed depth as guaranteed. Orders can cancel and hidden size is not included.
• Pooling raw dollar depth across years. Market price and size drift can turn a clock study into an era study.
• Mixing weekdays and weekends. Their activity and volatility cycles differ.
• Calling one hour the best time to trade. Order size, direction, venue and strategy determine which condition is useful.
• Ignoring daylight-saving changes. Local market events move by one UTC hour during parts of the year.
• Letting the average override the live book. A liquidation or news event can dominate the normal session pattern.

What this study cannot answer

• Displayed depth does not measure hidden liquidity or cancellation risk.
• Hourly medians do not estimate the slippage of a specific order size.
• The historical depth sample is Binance-focused and ends in March 2026.
• BTC, ETH and SOL share market regimes and are not independent experiments.
• Complete-day filtering excludes days with feed gaps and may underrepresent some disorderly periods.
• UTC-hour averages blur daylight-saving shifts and event-specific releases.
• The results describe conditions, not a tested best-time-to-trade strategy.