What is the most played game in the world?

It sounds like a question with a number attached. It is not, and the reason is worth more than the answer.

The short version

If you want one name: Roblox, at roughly 144.5 million daily active users.

Then it gets complicated, because the runners-up are measured in different units.

Those four numbers cannot be ranked against each other. Registered accounts is a lifetime total including everyone who quit in 2019. Daily actives is a headcount for one day. Peak concurrent is the single busiest moment. They differ by more than an order of magnitude in what they even describe.

Putting them in one sorted list, which nearly every "most played games" article does, produces a ranking of which company chose the most flattering metric.

The metric that matters, and who hides it

Daily active users is the honest number. It answers "how many humans played this today", which is the actual question.

Now the part that explains the whole mess:

Only Roblox, Activision Blizzard and a handful of Chinese publishers publish DAU. Almost nobody else does.

Everyone else releases registered accounts, or lifetime sales, or "players since launch", or nothing at all. Those figures are chosen because they only go up.

So the field is left with two substitutes:

Steam concurrency, which is real, live and precise, and covers only Steam. Valve exposes it through its own API, which is why Steam Charts and similar sites can be exact rather than estimated.

Estimates, from Newzoo and others, which model monthly actives across platforms. Genuinely useful, genuinely modelled, and often quoted as though measured.

Where to actually look

The rule that keeps you out of trouble: only one of these is a measurement, and it only covers PC games on one store. Everything else is a model or a marketing choice. That is not a criticism of the modellers, who are open about their method. It is a criticism of everyone who reprints their output without the word "estimated".

The three holes

Three enormous parts of the medium have no reliable player counts at all.

Mobile. The largest platform by revenue, at 107 billion dollars in 2026, and the one where numbers are least available. App store rankings show relative position, never absolute play.

Worlds inside platforms. A popular Roblox experience may have more daily players than a mid-size Steam hit. It has no Steam page, no store listing, no independent tracker. It exists inside somebody else's number.

The browser. Over a million games on itch.io, thousands from jams, all the vibe-coded work of the last two years. No counter exists. This is the hole we are building into, and the honest reason we can attempt it is that measuring it now costs one script tag rather than an SDK integration.

The answer, stated properly

Roblox is the most played game in the world, by the only metric anyone publishes honestly, and the gap is not close. 144.5 million people a day is a different category of thing from 1.36 million simultaneous players on the biggest game on Steam.

But notice what that sentence contains. It is true because Roblox chose to publish. If Epic published Fortnite DAU tomorrow, the ranking might survive or might not, and neither you nor we would have seen it coming.

A ranking that depends on who volunteers their numbers is not a measurement of the world. It is a measurement of disclosure policy.

That is the actual state of play in 2026, and it is why we are building an index that says, on every single figure, whether it was self-reported, measured, or inferred.

Sources

• Roblox daily active users, Fortnite registered accounts and concurrency, Minecraft daily players, as compiled in 2026 player-count reports by DemandSage and ShaneTheGamer
• Counter-Strike 2 peak concurrency for June 2026 and Steam ranking data via Steam Charts and ActivePlayer.io
• ActivePlayer.io methodology for concurrent, peak, DAU and MAU figures
• Newzoo for monthly active estimates and the 2026 mobile market figure of 107 billion dollars
• Steam Web API as the source of official live concurrency

Retrieved 4 August 2026. Where a figure is an estimate rather than a measurement, the table above says so. Numbers describing different units are deliberately not ranked against each other.