What time do signal channels post? 3,988 calls in UTC

We timestamped every signal in the index, as of August 2026. A quarter of all calls land in the six hours after midnight UTC, and the busiest hour carries under four times the traffic of the quietest, though one publisher supplies 48.35% of the sample.

What does the index say about publishing hours?

That the feed runs all day. Across 3,988 signals from 22 publishing channels, the busiest hour of the day is 13:00 UTC with 244 calls, or 6.12% of the sample, and the quietest is 23:00 UTC with 67 calls, or 1.68%. Every one of the twenty-four hours carries traffic. The gap between the loudest and the quietest is under four to one, which for a distribution across a full day is close to flat.

Here is the whole clock, counted in UTC on the publication timestamp of each message, in the index snapshot that ends on 17 August 2026.

A perfectly even feed would put 4.17% in every row. Thirteen of the twenty-four hours sit within a point of that figure. The market these channels call has no closing bell, and on this evidence most of the channels do not have one either.

Is there an hour when nothing arrives?

No, and the night is not even quiet. The six hours from 00:00 to 06:00 UTC carry 1,036 signals, which is 25.98% of the sample against the 25% that six hours of a flat day would produce. The twelve daylight hours from 06:00 to 18:00 UTC carry 59.0%, against 50% for a flat day.

So there is a daytime tilt, and it is about nine percentage points wide. That is the entire working-hours effect in the data.

For a subscriber, the practical version of that number is blunter. The window from midnight to 06:00 UTC is the small hours in London, the evening before in New York and the working morning in Singapore. Roughly one call in four lands inside it, so whichever of those clocks is yours, part of the feed arrives when you are not reading it.

Whose day is this histogram?

Mostly one publisher's. A single channel, Free Pump Crypto Signals, supplies 1,928 of the 3,988 signals, which is 48.35% of the sample. The top three channels supply 65.15% and the top ten supply 88.77%. Any aggregate here is heavily a description of a handful of feeds.

There is a second distortion underneath the first. Six clone links recorded in the index connect four channels: three of them carry virtually the same stream, with 334 of their 335 signals in common and the clone link recording the overlap as 100%, and a fourth shares 72 of its signals with that feed. Counting brands as independent publishers counts one editorial clock several times. Treating the reposting brands as duplicates removes 735 signals and leaves 3,253.

The useful test is to take the dominant channel out and see whether the shape survives. It largely does. Across the remaining 2,060 signals, the busiest hour is still 13:00 UTC at 6.36%, the night window falls from 25.98% to 21.46%, and the weekend share rises. The night share drops, the flatness stays.

Do channels take the weekend off?

Slightly, and less than a desk job would predict. Saturday and Sunday together carry 936 signals, or 23.47% of the sample, where an even week would put 28.57% into two days out of seven.

Monday is the loudest day of the week by a clear margin, and Saturday the lightest. Take the dominant channel out and the weekend share rises to 24.37%, so even the modest weekend dip is partly one publisher's habit.

Does any channel actually keep hours?

One does, and it is the smallest of the five publishers with enough volume to profile. CryptoNinjas Trading published 159 signals in the sample, 19.5% of them in the 09:00 UTC hour alone, and 80% of its output falls inside ten hours of the day. Five hours of the day contain nothing at all.

That is what a schedule looks like in this data: a peak more than three times the height of the market-wide one, and a block of the day that is empty instead of thin.

It is also the exception. No other publisher in the profiled set comes close to that shape, which is why the flat corpus-wide histogram should not be read as "channels post evenly around the clock by design". It is closer to the sum of several feeds that each ignore the clock in their own way.

What does a feed without hours look like?

The busiest publisher in the index spreads 80% of its calls across sixteen hours of the day and posts 30.81% of them in the six-hour night window, which is a higher night share than the corpus average it dominates.

The three middle rows are the visible part of the clone cluster, and their near-identical numbers are the point of including them: the same stream under three brand names produces the same clock three times. The fourth member is too small to profile here and shares 72 of its signals with them. A subscriber comparing those feeds side by side would be comparing one publishing schedule to itself.

Can you read a channel's location off its clock?

You cannot, and the temptation is worth naming because the numbers invite it. A feed that goes quiet from 20:00 to 01:00 UTC looks like somebody's evening, and a feed with a 06:00 UTC peak looks like somebody's morning. Neither is evidence.

Our index carries no usable location for any operator. The country field was never populated and is not part of this snapshot at all, so there is nothing to check an inference against, and the clone cluster shows why the inference is shaky in the first place: three brands with one histogram between them describe a publishing pipeline, and a pipeline's schedule says nothing about any person's day.

What does the publishing hour do to whoever executes it?

It sets the distance between the message and the trade. A call is written against the price on the screen at the moment it is sent, and the person acting on it reads it later by some amount, which the timestamp fixes but the post never mentions.

That delay is the thing the rest of our research keeps running into. In this snapshot, 1,958 of 3,927 replayed outcomes, 49.86%, are cancelled: price reached the take-profit zone before the published entry ever filled. Our earlier piece on that pattern, half the trades never happen, works from a 7 August 2026 cut of the index and breaks the rate down by publisher, because it is concentrated across the corpus and not uniform.

The timing data adds one line to that argument. A quarter of all calls are published into hours when a large part of the subscriber base is not reading, which puts the delay on those calls in hours instead of seconds. What a call looks like once price has moved away from its entry is set out in when to skip a crypto signal that has already moved.

What these timestamps cannot tell you

Less than the round numbers suggest, in four specific ways.

1. Coverage. Signals exist for 22 of the 47 channels in the index, and the other 25 contribute nothing. Across the 22 channels our feed assessment has reviewed, 19 carry no entries at all to parse: promo funnels (10), market commentary (6) and results-only posts (3). For most of the missing publishers there is no clock to compare, never mind a matching one.
2. Period. The formal span runs from November 2021 to 17 August 2026, but 90.37% of the signals were published in 2026 and 1,956 of them in July 2026 alone. August 2026 is itself a partial month. This is a summer-2026 sample with a long thin tail, not five years of history.
3. What a timestamp is. It records when a message appeared in a public channel. It says nothing about when the analysis was done, whether the post was scheduled, or whether an identical call went to a paid tier earlier.
4. Freshness. The last signal in this snapshot is dated 17 August 2026 at 03:01 UTC. Channels change habits, and a publisher that starts or stops posting will move these shares.

What a subscriber can check on a feed's own clock

Three things, all visible in a channel's public history without any opinion about the market.

1. The hour histogram of that one channel. The corpus-wide histogram is dominated by a single publisher, so the only clock that tells you anything is the one belonging to the feed you are reading.
2. Whether the quiet hours are quiet or empty. A feed with genuinely empty hours is behaving like an operation with a routine. A feed that is merely thin at 03:00 UTC will still send you calls at 03:00 UTC.
3. Whether its busiest hours are hours you are awake to read. This is arithmetic on two clocks and needs no view on the market. Nothing here recommends taking, avoiding or timing any trade.

A signal posted at 02:00 UTC is a set of price levels with a shelf life, and the hour it went out was chosen by the publisher rather than the reader.

Sources and method

The data. Every figure comes from one snapshot of the ChainRated signal index, whose last signal is dated 17 August 2026 at 03:01 UTC: 3,988 parsed signals from 22 publishing channels out of 47 indexed accounts, all of them public Telegram channels, spanning 9 November 2021 onward. Hours and weekdays are extracted from each message's publication timestamp in UTC, with no local-time conversion at any point.

Concentration. One publisher supplies 48.35% of the sample, the top three 65.15% and the top ten 88.77%. Six clone links connect four channels that republish a shared stream; dropping the repost duplicates leaves 3,253 signals. Where a figure changes materially once the dominant publisher is removed, both versions are printed above.

Replay outcomes. The cancelled count quoted here uses the same snapshot as the timing figures, which is a later cut than the 7 August 2026 one behind our earlier article on unfillable entries, so the two totals differ by design and should not be combined. How a published signal becomes a resolved outcome is on the methodology page.

What this does not establish. These are timestamps on public messages. Nothing here measures intent, alleges misconduct by any named channel, or identifies where any operator is located. Nothing here is a recommendation to take or avoid any trade.