Movement · risers and fallers
Movement among channels of 316,228–1,000,000 subscribers
4 entries in this band were measured to change size; the boards below are the 25 largest movements of each kind among them. A channel is placed in a band by its latest reading, so an entry that crossed a band edge during the window is listed under where it ended up.
Hours of history, not weeks. The register began recording subscriber counts on 5 August 2026. The median entry on these boards was watched across 16 hours, and every row states its own interval. Nothing on this page has been divided by time to produce a rate.
Biggest gains in this band
The largest measured increase among entries whose latest reading falls in 316,228–1,000,000. Segmenting by size is what makes this board answerable: on the whole-register board a channel of this size is outranked by a far larger one moving a fraction of a percent, every time.
| # | Entry | Change | As a percentage | First reading | Latest reading | Observed over |
|---|---|---|---|---|---|---|
| 1 | SOSO吃瓜 - 吃瓜TOP 🔥@sosochigua | +3 | +0.001% | 426,840 | 426,843 | 23 minutes |
Biggest losses in this band
The largest measured decrease in the same band. Telegram removes deleted and deactivated accounts in batches, so a fall is routinely housekeeping rather than an audience leaving; this board reports the measurement and draws no conclusion from it.
| # | Entry | Change | As a percentage | First reading | Latest reading | Observed over |
|---|---|---|---|---|---|---|
| 1 | 东南亚大事件@kanxia | −713 | −0.152% | 470,157 | 469,444 | 28 hours |
| 2 | BYIN Channel@byin_fun | −274 | −0.051% | 537,710 | 537,436 | 6.7 hours |
| 3 | 东南亚大事件@bx6666o | −251 | −0.048% | 518,700 | 518,449 | 14 hours |
Fastest growth in this band, proportionally
The same movements as a share of the channel’s own size. Inside one half-decade of size this is a much fairer comparison than it is across the whole register — every entry here is within a factor of 3.16 of every other, so a percentage means roughly the same amount of work throughout the board.
| # | Entry | Change | In subscribers | First reading | Latest reading | Observed over |
|---|---|---|---|---|---|---|
| 1 | SOSO吃瓜 - 吃瓜TOP 🔥@sosochigua | +0.001% | +3 subscribers | 426,840 | 426,843 | 23 minutes |
Steepest decline in this band, proportionally
The same population and the same floor, in the other direction. The absolute figure is printed beside every percentage so a proportional fall is never read without the number of people it stands for.
| # | Entry | Change | In subscribers | First reading | Latest reading | Observed over |
|---|---|---|---|---|---|---|
| 1 | 东南亚大事件@kanxia | −0.152% | −713 subscribers | 470,157 | 469,444 | 28 hours |
| 2 | BYIN Channel@byin_fun | −0.051% | −274 subscribers | 537,710 | 537,436 | 6.7 hours |
| 3 | 东南亚大事件@bx6666o | −0.048% | −251 subscribers | 518,700 | 518,449 | 14 hours |
Movement in every size band
What was measured, and what was held back
The figures below describe the whole register, not this band alone: they are the one measurement pass every board on this site is cut from, and stating them per band would invite the eight pages to be added up into a total they do not make.
| Entries read twice or more | 662 | On the t.me/name profile page, inside the window. An entry read once cannot appear on any board here. |
|---|---|---|
| Window | 28 Sept 2026, 10:16 → 30 Sept 2026, 10:00 | The oldest and newest readings behind these boards, UTC. This is the whole depth of the register’s subscriber history, not a slice of a longer one. |
| Boards computed at | 30 Sept 2026, 10:07 | UTC, and it is this figure rather than the clock that says when the numbers below were true. Every one of them comes from a single pass over the window, recomputed on a twenty-minute cycle and held briefly in memory here, so a page opened now can be showing a pass taken a little before that cycle suggests — the time above is the one that is measured. The pass being a single statement is what makes the counts in this table the denominators of the boards above them, rather than a second, later measurement of a slightly different set of channels. |
| Median entry observed over | 16 hours | Half the entries below were watched for less than this. Nothing on this page is divided by it to make a rate. |
| Went up | 149 | Together 21,984 subscribers gained. |
| Went down | 248 | Together 23,377 subscribers lost. The register moved down by 1,393 across every entry on these boards. |
| Read twice, did not move | 178 | A measured flat line, not an absence of data. Snapshots are written on change, so these are entries whose second reading agreed with the first. |
| Held back — inside the rounding | 36 | The change is no bigger than the rounding step its two readings could carry between them. Withheld rather than published as a movement. |
| Held back — doubled or halved | 51 | A change of that size across hours is a misread far more often than it is news. Nothing here decides which of the two readings was wrong; it declines to subtract them. |
Why the percentage boards have a floor. A channel going from 3 subscribers to 6 has grown 100%, and a board that ranked on that would be a list of the smallest entries in the register, refreshed by whoever gained a friend. Both readings must be at or above 1,000 for an entry to carry a percentage here — the same floor the published engagement baselines use for eligibility, so the two are one number rather than two house rules. Absolute change has no floor: it needs none, because one subscriber is one subscriber whatever the size of the channel it joined.
Other ways in
Movement across the whole register, or every entry in this band listed by size with the measured engagement baseline for the band. Each board shows at most 25 rows and does not page: a leaderboard is a top of something.