Twitch trends

A ranked view of what our sampling sees on Twitch: which games hold the largest audiences, how many channels are already sharing each one, which channels drew the most watch time on the last complete day, and which added the most followers. The number that matters most on this page is not the biggest audience, it is viewers per channel - a huge room split a thousand ways is a worse place to go live than a small one split ten.

Channels tracked

1,426,377

1,293,078 with a measured viewership

Audience measured

25.5M

Sum of average concurrent viewers

Categories seen

15,216

Top 24 shown below

Rollup day

2026-10-29

Source of the two channel boards

Twitch categories by audience and by room

Audience is the sum of the average concurrent viewership of every channel we have measured in the category. Room is that pool divided by the channels sharing it, and it is the column that answers whether a new stream can be found there.

Ranked on the size of the viewer pool.

Twitch categories ranked by measured audience and by viewers per channel
CategoryAudienceChannelsViewers per channelRoom
Just ChattingHealthy ratio of viewers to channels. Worth testing a stream in.4.3M16.8%44,01039,904 measured107
Counter-StrikeHealthy ratio of viewers to channels. Worth testing a stream in.1.2M4.8%14,15410,222 measured119.6
League of LegendsHealthy ratio of viewers to channels. Worth testing a stream in.971.7K3.8%14,67611,235 measured86.5
VALORANTContested. Expect to sit low on the listing without an existing audience.742.5K2.9%22,83219,076 measured38.9
SlotsAudience heavily outnumbers the channels serving it. Room to be seen.737.4K2.9%1,4951,391 measured530.1
Grand Theft Auto VHealthy ratio of viewers to channels. Worth testing a stream in.688.8K2.7%16,03212,850 measured53.6
Dota 2Audience heavily outnumbers the channels serving it. Room to be seen.679K2.7%6,9844,192 measured162
FortniteContested. Expect to sit low on the listing without an existing audience.665.9K2.6%37,26032,484 measured20.5
MinecraftContested. Expect to sit low on the listing without an existing audience.651.3K2.6%28,36524,385 measured26.7
IRLAudience heavily outnumbers the channels serving it. Room to be seen.613.6K2.4%3,6893,265 measured187.9
World of WarcraftHealthy ratio of viewers to channels. Worth testing a stream in.405.1K1.6%8,4374,244 measured95.4
Call of Duty: Modern Warfare 4Contested. Expect to sit low on the listing without an existing audience.362.5K1.4%16,49514,321 measured25.3
Special EventsAudience heavily outnumbers the channels serving it. Room to be seen.341.6K1.3%1,6511,281 measured266.7
EA Sports FC 27Contested. Expect to sit low on the listing without an existing audience.309.3K1.2%12,0708,475 measured36.5
DJsAudience heavily outnumbers the channels serving it. Room to be seen.295.3K1.2%3,1252,384 measured123.9
OverwatchContested. Expect to sit low on the listing without an existing audience.285.2K1.1%14,25911,353 measured25.1
Dead by DaylightContested. Expect to sit low on the listing without an existing audience.270.1K1.1%15,69013,029 measured20.7
SportsHealthy ratio of viewers to channels. Worth testing a stream in.239.6K0.9%2,6532,340 measured102.4
DeadlockContested. Expect to sit low on the listing without an existing audience.232.7K0.9%9,3246,296 measured37
Apex LegendsContested. Expect to sit low on the listing without an existing audience.232.1K0.9%12,4689,003 measured25.8
AION 2Healthy ratio of viewers to channels. Worth testing a stream in.228.7K0.9%5,7284,992 measured45.8
MusicHealthy ratio of viewers to channels. Worth testing a stream in.221.9K0.9%5,4523,277 measured67.7
ROBLOXCrowded. Most channels here average single-digit viewers.201.8K0.8%24,89522,637 measured8.9
ASMRAudience heavily outnumbers the channels serving it. Room to be seen.197.9K0.8%1,247886 measured223.4

Channels is the full roster we track in the category; the audience and the ratio use only the ones with a viewership on record. A channel enters our sample only when it is on air as a sweep runs, so the roster is always the larger number.

Most watched Twitch channels

Hours watched on 2026-10-29: average concurrent viewers multiplied by time on air, which measures attention rather than reach.

  1. 1StRoGoCounter-Strike - 2.3M followers292.2K h watched30K average viewers
  2. 2MastuMinecraft - 1.8M followers290.9K h watched16.6K average viewers
  3. 3PGLCounter-Strike - 2M followers262K h watched26.2K average viewers
  4. 4CaedrelLeague of Legends - 1.7M followers245.8K h watched57.8K average viewers
  5. 5TheBurntPeanutAge of Empires II - 2.6M followers235.8K h watched39.3K average viewers
  6. 6加藤純一うん〇ちゃんJust Chatting - 1.2M followers210.1K h watched27.1K average viewers
  7. 7CDawgJust Chatting - 1.5M followers176.6K h watched10.9K average viewers
  8. 8xQcGrand Theft Auto V - 12.6M followers165.9K h watched23.7K average viewers
  9. 9afoninjeDota 2 - 154.5K followers165.6K h watched18.4K average viewers
  10. 10fissure_dota_ruDota 2 - 556K followers153.4K h watched16.1K average viewers
  11. 11GorgcDota 2 - 788.6K followers126.9K h watched14.1K average viewers
  12. 12stylishnoob4AION 2 - 1.2M followers126.4K h watched13K average viewers

Fastest growing Twitch channels

Followers added on 2026-10-29, ranked on the absolute number. The rate is shown beside it, because on a platform where the largest channels add thousands a day the two say different things.

  1. 1skulaa_enjoyerCategory unknown - 133.1K followers+67.1K followers101.73% in a day
  2. 2a1gyrcezCategory unknown - 55.2K followers+11.7K followers26.75% in a day
  3. 3LimmigreParisienCategory unknown - 240.4K followers+10.8K followers4.69% in a day
  4. 4ezekielthe1stCategory unknown - 80.9K followers+10.3K followers14.55% in a day
  5. 5xdityyyCategory unknown - 44.4K followers+10.1K followers29.48% in a day
  6. 6YingCategory unknown - 95.5K followers+9,867 followers11.52% in a day
  7. 7evelynomaderaIRL - 686.5K followers+8,743 followers1.29% in a day
  8. 8MastuMinecraft - 1.8M followers+7,522 followers0.42% in a day
  9. 9IShowSpeedCategory unknown - 6.3M followers+6,841 followers0.11% in a day
  10. 10CaedrelLeague of Legends - 1.7M followers+5,475 followers0.31% in a day
  11. 11PeakFighterrCategory unknown - 61.7K followers+5,418 followers9.62% in a day
  12. 12LenovoLegionCategory unknown - 166.8K followers+5,378 followers3.33% in a day

When Twitch is least contested

Total viewers divided by live channels, by hour of day in UTC, across every channel our sweep caught live in the last 3 days. Taller is emptier.

0
3
6
9
12
15
18
21

What the board says

  • Just Chatting is the largest audience pool in our Twitch catalog: 4,268,680 average concurrent viewers spread across 39,904 measured channels, which is 107 viewers per channel.
  • The least contested category on the board is Slots, at 530.1 viewers for every channel we measure in it - 5x the ratio in Just Chatting.
  • StRoGo drew the most watch time of any tracked Twitch channel on 2026-10-29: 292,173 hours watched.
  • skulaa_enjoyer added the most followers of any tracked channel that day, 67,114 (101.73%).
  • Across the whole platform the emptiest hour of the day is 8 PM UTC, where each live channel in our sample has 164.3 viewers to draw on.

Where these numbers come from

  • The category board is built from our own sampling, not from a Twitch endpoint. "Viewers" is the sum of the average concurrent viewership of every channel we have measured in that category, so it is the size of the audience pool over our measurement window rather than a headcount at this instant.
  • Channels with no viewership on record are counted in the channel column and excluded from every viewer figure. A channel only enters our sample if it is on air when a sweep runs, so the roster is always larger than the measured set.
  • The two channel boards come from the nightly rollup for 2026-10-29, which is the most recent complete day we hold.
  • The hour curve covers the last 3 days of viewer sampling across every Twitch channel our sweep caught live, in UTC.

What we actually measure on Twitch

Every figure on this page is our own measurement, not a Twitch endpoint. Our directory engine reads Twitch's public live listing every 15 minutes, appends the concurrent viewer count for each live channel, and rolls the readings into one row per channel per day. As of 2 September 2026 that gives 66,493 active Twitch channels in the catalog, 41,506 of them with a measured average concurrent viewership, across 2,018 games and 37 languages.

The audience column on the game board is a sum of averages: for every channel we have measured in that game, its average concurrent viewership, added up. That makes it the size of the audience pool over our measurement window rather than a headcount at this instant, and it is the honest way to size a room from sampled data. Channels we have never caught live are counted in the channel column and excluded from every viewer figure, which is why the roster is always larger than the measured set.

Viewers per channel is the column to read second and act on first. Twitch's biggest game is almost always the most crowded one, so the pool being large tells a new channel very little; the ratio between the pool and the number of channels sharing it tells them whether they will be seen. The board prints both and states which is which, and the verdict on each row is built from the ratio rather than from the rank.

The two channel boards come from the most recent completed nightly rollup, so they answer "the last full day we hold" rather than "right now". Most watched is ranked on hours watched, which is average concurrent viewers multiplied by time on air, and it is the closest thing to a measure of attention that sampling can produce. Fastest growing is ranked on followers added that day; the percentage is shown next to it, because on a platform where the biggest channels add thousands a day the absolute number and the rate say different things.

Nothing here is a Twitch statistic and nothing here is complete. Our sweep browses games rather than enumerating every live channel, so the tail is undercounted, and a channel that streams between two sweeps can be missing from a day entirely. Shares and ratios survive that; absolute counts are ours and are labelled as ours.

How to read the board

Audience is the sum of the average concurrent viewership of every measured channel in the game. Channels is the full roster, measured is how many of them have a viewer figure at all, and viewers per channel divides the first by the third.

The room bar is relative to the widest game on the board, not to an absolute threshold, because the same viewers-per-channel figure means different things on Twitch than it does elsewhere. A full bar is the roomiest game shown; an empty one is the most contested.

The hour curve underneath is the whole platform over the last three days of sampling, in UTC. It is the same curve the Twitch stream time tool uses to rank hours a channel has never tried.

Catalog figures quoted in this section were measured on 2 September 2026 across 66,493 active Twitch channels, 41,506 of which carried a measured average concurrent viewership.

What this board does not do

It does not tell you what is trending in the last hour. Our game figures are averages over a sampling window and the channel boards are yesterday's rollup, so this is a picture of where attention has been sitting, not a live ticker. It also cannot see channels it has never caught broadcasting, which biases every count towards channels that stream often and long. And a roomy game is an opportunity, not a promise: the ratio says you can be seen there, not that the audience wants what you make.

Frequently asked questions

Is this Twitch's own data?

No. PlayerSells is not affiliated with Twitch. Every number here comes from our own sampling of Twitch's public live listing, taken every 15 minutes and rolled up nightly. That is why we can show a viewers-per-channel ratio that Twitch does not publish, and why our absolute counts are smaller than the platform's own.

Which game should I stream in?

Read the viewers-per-channel column, not the audience column. The biggest game is nearly always the most crowded, so an unknown channel is more likely to be found in a game where the audience outnumbers the channels serving it. The verdict on each row is built from exactly that ratio.

Why is a channel I know missing from these boards?

Because it did not appear in the last complete rollup day, or because we have never caught it live. A channel only enters our sample if it is broadcasting when a sweep runs, so quiet weeks and short streams are under-represented. Nothing is filtered out for editorial reasons; adult-labelled channels are the one exception and are excluded everywhere on this site.

How often does this update?

The catalog figures move with our sampling and are cached for half an hour. The two channel boards move once a day, when the nightly rollup completes, and the page prints the day they came from.

What does hours watched mean?

Average concurrent viewers multiplied by hours on air, over the rollup day. One viewer watching for ten hours and ten viewers watching for one hour both count as ten hours watched, which is why it measures attention rather than reach.

PlayerSells is not affiliated with Twitch. Every figure on this page is our own measurement, sampled from Twitch’s public live listing every 15 minutes and rolled up nightly. Twitch publishes no viewer history, so nothing here is a Twitch statistic and nothing here should be read as one.