Roblox Charts
Live player counts for 874 qualifying Roblox games, measured hourly by us and ranked four different ways.
Hourly player-count collection began , which is 9.2 hours ago across 4 completed runs. We have no readings from before that date and do not estimate any, so nothing here describes a period we did not measure. Charts covering longer periods activate on their own as the readings build up.
The four charts
Busiest right now
- 1
1,134,746 playing
- 2
616,534 playing
- 3
362,889 playing
- 4
303,233 playing
- 5
300,628 playing
How these charts are built
Most Roblox chart sites show you a number and ask you to trust it. These pages try to do the opposite: show the number, and show enough of its construction that you could argue with it.
Once an hour a job asks the Roblox games API for the live concurrent player count of every game in our catalogue, and writes one row per game per run. That row is never updated. Everything on these charts — a 24-hour change, a trending score, a rank — is computed from those raw rows when you load the page, not stored as a pre-baked column. That matters more than it sounds: if we stored a “24h change” figure, its definition would be frozen at the moment it was written, and every later improvement to how we handle outliers, baselines or missing readings would only apply going forwards. Deriving it on read means every observation we have ever taken stays usable under whatever the current definition is.
The baseline for any window is the newest reading at or before the edge of that window, never the oldest reading inside it. Taking it from inside would make a change figure shrink as more data arrived, which reads as a game losing momentum when in fact nothing happened but us collecting more. Each row also carries the true age of the baseline it used, which is what lets these pages refuse to call a four-hour comparison a daily one.
When Roblox fails to answer for a game, the failure is recorded in a separate table and no snapshot row is written. A gap in a chart is therefore a real gap, and the player-count graphs break their line across it rather than drawing a straight segment that never happened.
What these charts still cannot tell you
Three things worth knowing before you read too much into a row.
Roblox concurrency has a violent daily cycle. Nearly every game rises through the American afternoon and falls overnight, so over a short window most games move together. A raw percentage over a few hours is mostly telling you the time of day. The movement charts print what the median qualifying game did over the identical window next to every row, so you can see the tide and judge the swimmer; the trending score removes the shared component outright by scoring each game against the distribution of what every other game did.
An event spike is not growth. A game that doubles because it shipped an update or got a video will decay back over the following days. Nothing on these pages can tell the difference yet — separating a durable rise from a spike needs weeks of history behind each game, and we do not have weeks.
Highest tracked is not all-time. Where a page mentions a peak it means the highest value we have recorded since collection started, and it says so. We hold no data at all from before that date and will not pretend otherwise, however tempting a round number would look.
Questions about these charts
From our own hourly sweep of the Roblox games API. Once an hour we ask Roblox for the live concurrent player count of every game in our catalogue and store the answer as a dated reading. Nothing is scraped from another chart site, nothing is estimated, and no reading is ever edited after the fact. When Roblox does not answer for a game, we record that it failed rather than writing a zero — a failed request and an empty game must never look the same on a chart.
It is a game's position within the set of games we track, in a single hourly sweep — not a rank published by Roblox. Roblox does not expose a global ranking at all, so anyone presenting one as though it came from Roblox has computed it themselves against whatever list they happen to hold. We show ours as "RBLXDB #12/1197" so the denominator is always attached: a rank without the size of the field it was measured in is not information.
Because there is not. Hourly collection is new, and a chart covering seven days cannot be built from a few hours of readings. Rather than stretching what we have across a longer axis and hoping nobody checks, the pages state exactly how much history exists and what it can support. The missing charts fill themselves in as the readings accumulate — no backfill, no manual step.
Because percentages on tiny numbers are meaningless. A game going from 2 players to 20 is up 900% and would top any growth chart, while telling you nothing. Concurrency also carries natural random jitter of roughly the square root of the count, so a 20-player game wobbles by about 22% on its own with nothing happening. We require 500 concurrent players at both ends of the window, where that jitter is down around 4.5% and a real move stands clear of it.
Readings are collected once an hour and the pages regenerate on a half-hour cycle, so what you see is at most about ninety minutes behind the live count. Each page states the age of the reading it is showing.
No, and no chart honestly can — tens of millions of places exist. We track the games that hold a meaningful live player base, which is the only population where concurrent-player figures mean anything. Every page tells you how many games were in the set it ranked.
Elsewhere on rblxdb
Best Roblox games by genre → · Live Roblox player stats → · The full Roblox game database → · Working game codes → · How we verify what we publish →