It’s 2 a.m. in New York on Saturday, August 22, 2026, and Ether has just fallen almost 3% in an hour, from about $2,514 to $2,440. Open a price app at that moment and it tells you ETH is up 3.02%. Open a daily chart and the candle says ETH is down 3.03%.
Nobody made a mistake. The two numbers answer different questions. Knowing which question you’re asking is one of those small, unglamorous edges that quietly saves you from bad entries, false alarms and a lot of confusion.
Two clocks, one price
Think of the two numbers as two clocks that were started at different times.
The rolling 24h change always looks back exactly 24 hours. At 06:00 it measures from 06:00 yesterday; an hour later, from 07:00 yesterday. It’s a window that slides forward with the clock, and nothing about the coin has to change for its starting point to move.
The daily candle is a clock reset once a day, at 00:00 UTC. Binance, Coinbase and Kraken all open their daily candles at midnight UTC. From then until the close, the candle’s open stays where it is, and the day’s change is measured from it. It’s the same candle you see on a daily chart.
24h change (rolling)
+3.02%
since Fri 06:00 UTC
Daily candle change
−3.03%
since the 00:00 UTC open
Here’s what happened. On Friday evening, between 21:00 and 22:00 UTC, ETH jumped about 3%. When Saturday’s daily candle opened at midnight, that rally was already history: it belongs to Friday’s candle. But it was still inside the 24-hour window.
So when ETH dropped at 05:00, the daily candle turned clearly red while the 24h number was still carrying Friday evening’s gains. From 02:00 UTC onward, Saturday’s candle never turned green again. It closed the day down 3.72%. The 24h number, meanwhile, stayed green for most of the day.
The moment yesterday leaves the window
Now move the slider to 22:00 UTC on Saturday. In that hour ETH slipped about 1%, from $2,432 to $2,408. The 24h change fell almost four points, from −0.38% to −4.21%.
The extra drop had nothing to do with Saturday’s trading. At 22:00, Friday’s rally candle slid out of the window. The starting point jumped 3% higher, and the 24h change fell with it. Economists call this a base effect: a change in the starting point, not in the thing being measured.
A rolling 24h number can move a lot while the price stands still, because its starting point keeps moving.
It works in both directions. A coin can “rally” 5% on a 24h list at the moment yesterday’s crash leaves the window, without a single new buyer showing up.
The daily candle has a blind spot of its own. At 01:00 UTC it has seen one hour of trading, so a +2% reading is one hour’s move, not a day’s. As the day goes on, the two numbers cover more and more of the same hours. Just before the close they nearly match, and at the close they’re identical.
How often do they disagree? We checked four months of data
Saturday’s ETH wasn’t a fluke. We took 131 of Binance’s most-traded USDT pairs and compared the two numbers at the end of every hour from July 17 to November 13, 2026: about 371,000 snapshots.
- 12.8%
- 12.8%
- of hourly snapshots pointed in opposite directions (beyond ±0.5% each)
- 23.7%
- 23.7%
- were 3 percentage points or more apart
- 2.8 / 10
- 2.8 / 10
- top-10 gainers shared by both lists one hour after the daily open
The time of day matters enormously. Right after the daily open, the daily candle is brand new while the 24h window is still full of yesterday. Between 02:00 and 07:00 UTC, the two numbers pointed opposite ways in roughly one snapshot in five.
Top-10 gainers that appear on both lists
Average by hour of the UTC day (out of 10)
Read the chart from the left. One hour after the open, the two top-10 lists share fewer than 3 coins on average. They only converge late in the day. If you scan a 24h top-gainers list in the first hours after the UTC open (evening in the Americas, morning in Asia), more than half of it usually isn’t among today’s top 10 daily candles. It’s mostly yesterday’s story.
When the daily candle closes where you live
The daily candle runs on UTC, not on your clock. For most of the Americas, the crypto “day” turns over in the late afternoon or evening. In Asia, it turns over in the morning.
Your daily close
The daily candle closes every day at 00:00 UTC.
| City | Local time of the 00:00 UTC close (standard time) | During daylight saving |
|---|---|---|
| Los Angeles | 4:00 PM | 5:00 PM |
| New York | 7:00 PM | 8:00 PM |
| São Paulo | 9:00 PM | — |
| London | 12:00 AM (midnight) | 1:00 AM |
| Berlin, Paris | 1:00 AM | 2:00 AM |
| Istanbul | 3:00 AM | — |
| Dubai | 4:00 AM | — |
| Mumbai | 5:30 AM | — |
| Singapore, Hong Kong | 8:00 AM | — |
| Tokyo, Seoul | 9:00 AM | — |
| Sydney | 10:00 AM | 11:00 AM |
— means the city doesn’t use daylight saving time.
Daylight saving moves your local close by an hour twice a year; UTC never moves. If you trade daily closes, set your reminders in UTC, or let a tool count down to the close for you.
Which number should you use?
Neither number is better. Each one is right for a different question:
| Your question | Look at | Why |
|---|---|---|
| Which coins moved most over the past day? | 24h change | It always covers a full day, whatever the time. Good for a broad scan. |
| Is today’s candle breaking out? | Daily candle, current | It matches the candle on your chart and the level everyone will see at the close. |
| Did the move hold into the close? | Daily candle, last closed | A closed candle can’t change any more. The forming one can. |
| What is moving right now? | 1h or 15m candle, current | Both daily numbers are too slow to show what happened in the last hour. |
| Is today’s volume unusual? | Candle volume vs. its own average | Comparing a candle with recent candles of the same size removes the time-of-day bias. |
The same split applies to volume. “24h volume” is a sliding window, while daily candle volume counts only what has traded since 00:00 UTC. Early in the UTC day, daily volume always looks thin. That isn’t a lack of interest; the day has barely started.
Four costly mix-ups
1. Chasing yesterday’s move
A coin tops the 24h gainers list at 03:00 UTC. It looks hot, but the whole move may have happened 20 hours ago, and today’s candle may already be red. Before acting on a 24h list, check today’s candle and the last few hourly candles. A 24h number is a rear-view mirror.
2. Panicking at a phantom drop
As the ETH example showed, a 24h number can fall several points in an hour while the price barely moves, just because a big candle from yesterday left the window. If the price chart looks calm but the 24h change just collapsed, look 24 hours back before you react.
3. Treating the forming candle as finished
A daily candle that is up 8% at 14:00 UTC is a candle in progress. Long wicks are what’s left of moves that didn’t hold until the close. If your idea depends on how the candle closes, wait for the close or watch the last closed candle instead.
4. Comparing numbers from different apps
Two apps can show different 24h numbers for the same coin. They may use different exchanges, average several of them, or measure from a different start time. Some exchange apps even let you choose between “last 24 hours” and a fixed daily start. Before comparing, check what each number is measured from.
Seeing both numbers in Coinlyzer
Coinlyzer is built around candles, so it keeps these measures apart instead of leaving you to guess which one you’re looking at:
- 24h change (rolling): the familiar number, built from 5-minute candles.
- 1-day candle change, current: today’s candle as it forms, measured from the 00:00 UTC open.
- 1-day candle change, last closed: yesterday’s finished candle.
- The same split for every timeframe from 5 minutes to 1 week, across Binance, Kraken and Coinbase.
Put them side by side and the disagreements become something to look into rather than a source of confusion. Two screens to start with:
Up on 24h, red today
The 24h number still looks strong, but today’s candle has turned. The rally is in the past; worth knowing before you act on a gainers list.
All exchanges · 24h volume at least $1M · 24h change (rolling) above 3% AND 1 day candle change (current) below -1%
Green today, flat on 24h
Today’s candle is up, yet over 24 hours the coin is roughly flat, often because it is recovering a late drop from yesterday. The 24h lists haven’t caught up.
All exchanges · 24h volume at least $1M · 1 day candle change (current) above 4% AND 24h change (rolling) below 1%
Add an alert to either screen and Coinlyzer messages you on Telegram or your phone when a new pair starts matching: at most once per candle, not every second. See everything the screener measures.
Frequently asked questions
Is the 24h change wrong?
No. It accurately measures the last 24 hours. It answers a different question from the daily candle, and its starting point moves forward every minute, so it can change even when the price doesn’t.
What time does the daily candle close in crypto?
On Binance, Coinbase and Kraken, daily candles open and close at 00:00 UTC. That is 7:00 PM in New York and 4:00 PM in Los Angeles during standard time, and an hour later during daylight saving time.
Why do two apps show a different 24h change for the same coin?
They may use prices from different exchanges, average several exchanges, refresh at different moments, or measure from a different starting time. Some exchange apps let you choose between the last 24 hours and a fixed daily start. Check what each number is measured from before comparing them.
Is the daily candle change measured from the open or from yesterday’s close?
A candle’s change compares its close with its own open. Because crypto trades around the clock, a daily candle usually opens at or very near the previous day’s close, so the two methods almost always agree within a fraction of a percent.
Does the same apply to 24h volume?
Yes. 24h volume is a sliding 24-hour window, while daily candle volume counts only what has traded since 00:00 UTC. Early in the UTC day, daily volume always looks small simply because the day has just started.
Should I set alerts on 24h change or on candles?
Candle-based alerts are easier to reason about: they fire when a candle crosses your level, and in Coinlyzer at most once per candle. An alert on a rolling number can fire just because an old candle left the window, with no new price movement.
Coinlyzer shows market data only. Nothing in this article is investment advice or a recommendation to buy or sell anything. Past price data doesn’t predict future prices. Figures are from Binance’s public market data and were checked when the article was updated.