How Many People Actually Play That MMO? How We Count What Publishers Won’t Say

Every MMO player eventually types the same question into a search bar: "is this game dead?" And the results are always the same. A Reddit thread where one person says the starting zones are empty and another swears their server is packed. A YouTube video with a red arrow on the thumbnail. A forum post from 2019. Anecdotes, vibes, and outrage — everything except a number.

There's a reason for that. The companies that run these games know exactly how many people are playing, down to the minute, and most of them will never tell you. This post explains why that is, what can actually be measured from the outside, and how we turn those measurements into the population estimates you see on every game page of this site.

The numbers publishers give you, and why they're mostly useless

World of Warcraft reported subscriber numbers every quarter for over a decade. The last figure Blizzard ever published was 5.5 million, in 2015. After that: silence, replaced by "engagement is up year-over-year."

That pattern repeats across the industry, and the logic is simple: player counts only get published when they're flattering. When a company does share a number, it's usually engineered to be as large as possible while remaining technically true:

• "Registered accounts" counts everyone who ever made an account, including the person who quit during the tutorial in 2017.
• "Copies sold" tells you about launch marketing, not whether anyone is still logged in.
• "Monthly active users" counts the lapsed player who logged in once to claim a free mount.

None of these answer the question you're actually asking, which is: if I start playing this game tonight, will there be people in it?

What can actually be measured

You can't see inside a game company's dashboard, but live games leak activity everywhere. Some of those leaks are surprisingly high quality.

Steam concurrent players are the gold standard. Valve publishes real concurrent player counts through an official API — not estimates, actual logged-in players, updated continuously. When a game is on Steam, this is the closest thing to ground truth that exists outside the studio. The catch: plenty of MMOs aren't on Steam, or split their population across Steam and a native launcher. WoW isn't there at all. Final Fantasy XIV is there, but a large share of its players patch through Square Enix's own client.

Discord presence fills much of that gap. Nearly every living MMO has an official Discord server, and those servers expose how many members are online right now. People idle in Discord while they play, organize raids there, and drift away when they quit. It's a fuzzier signal than Steam, but it tracks the shape of a game's real activity remarkably well — and it works for games that never touch Steam.

Reddit's "here now" counts and Twitch viewership round out the live signals. Twitch in particular needs careful handling: it measures how watchable a game is, not just how played it is, and it spikes hard during drops campaigns and streamer events.

Then there are the slow signals — total Discord and Reddit membership, YouTube subscribers, Steam review counts, app store ratings. These don't tell you who's online tonight. They tell you how large the community around a game has ever been, and how fast it's still accumulating.

How we turn signals into an estimate

Every day, we collect all of the above for every game we track — around a dozen distinct metrics per game. Then we do two things with them.

First, we compute two scores per game: an activity score built from the "right now" signals (Steam concurrents, Discord online, Reddit online, Twitch viewers) and a community score built from the lifetime ones. The signals aren't treated equally. Steam's official concurrent count carries the most weight because it's real data; an app store rating count carries a twentieth of that, because it's the weakest proxy on the list. When a source has a gap — an API hiccup, a deleted Discord widget — we carry the last known value forward rather than letting a missing data point masquerade as a population collapse.

Second, we produce the estimated daily players figure you see on each game page. This step reconciles the scores with context that raw feeds can't capture: whether a game's audience lives mostly off Steam, whether it's regional, whether its Discord is unusually large or small for its genre. Every estimate comes with a written rationale on the game's page, along with the full history of its social signals, so you can see the trajectory rather than a single number.

We want to be precise about what this is: an estimate, not a census. No one outside the studio has a census. What we have is a method that's consistent, documented, and applied identically to over ten thousand games, every single day.

Where we're wrong, and by how much

Honesty about error bars is the difference between measurement and marketing, so here are ours.

We most likely undercount regional titles whose communities don't live on the platforms we read. A Korean MMO whose players organize on Naver cafés, or a Chinese one living on QQ, will look quieter in our data than it is in reality. We likely overcount highly watchable games during Twitch drop events, though weighting Twitch lowest among live signals blunts most of that. And private or emulator servers are invisible to us entirely.

Here's the important part: because these biases are stable, they mostly cancel out where it matters. Our absolute numbers carry uncertainty, but a game's trend — growing, stable, bleeding — and its rank relative to similar games are robust, because we're measuring every game with the same ruler every day. If a game drops 30% in our data over two months, that decline is real even if the starting number was imperfect.

How to actually use these numbers

"Is it dead?" is the wrong question. Better questions, in order:

1. Is there enough population for the way I play? Five thousand daily players in a niche sandbox like Project Gorgon or Wurm Online is a healthy, functioning world. Five thousand in a dungeon-queue themepark means you'll never see a queue pop. Population needs context — the genre and design of the game set the bar.
2. Which way is it trending? A modest population that's been stable for three years beats a bigger one in freefall. That's what the trending page is for.
3. How does it stack against the alternative you're considering? Every pair of games on the site has a comparison page showing their signals side by side.

An MMO isn't a $40 impulse purchase — it's a place you're going to live for a while. You wouldn't move to a town without checking whether people still live there. That's the entire reason this site exists: when publishers won't tell you who's actually playing what, measure it.

Start with the full standings and see for yourself.