Why gaming numbers are hard to compare
Few areas of culture produce as many large, confident figures as video games. Reports often describe billions of players, record concurrent users or audiences that rival traditional sport. Those figures are not necessarily wrong, but they are rarely comparable. Each one is typically the answer to a different question, asked by a different organisation for a different purpose, and the differences are usually buried in a footnote.
This article is about reading those claims. It does not offer its own statistics. It explains what the common measures tend to count, where they stretch, and what a careful reader can reasonably conclude about online gaming as a whole.
What “a gamer” can mean
The first problem is definition. A “gamer” might be someone who has played any game on any device in the past year, including a puzzle app on a phone during a commute. It might be someone who plays several times a week. It might be someone who identifies with the label. These are very different groups, and the first is typically far larger than the last.
Survey-based estimates depend heavily on which definition is used and how the question is worded. A person asked “do you play video games?” may answer differently from one asked “did you play any game on a phone, console or computer in the past month?” Many people who play casually do not think of themselves as gamers and may answer no to the shorter question. This is a general feature of how question wording shapes results, and gaming is a clear example.
Accounts, players and people
Platform-reported numbers add another layer. A publisher may report registered accounts, monthly active accounts, daily active accounts or peak concurrent users. Each has a legitimate use, and none is the same as the number of distinct people.
- Registered accounts count everyone who ever signed up, including people who tried a game once and left.
- Active accounts count accounts that did something within a time window, but the window (a day, a month) is chosen by the reporting company.
- Peak concurrent users count the most accounts online at one moment, which says little about how many people play over a month.
- Distinct people are hard to count at all, because one person may own several accounts and several people may share one.
This is a general limit of platform data: it is collected for the platform’s own operations and advertising, and it describes accounts and sessions rather than people. Automated or duplicate accounts can also inflate totals, a problem discussed more broadly under fake accounts.
Free-to-play, mobile and the shape of the audience
Much of the growth in online play is often attributed to mobile devices and to games that cost nothing to start. That shifts who is counted. A free game can attract very large numbers of short, casual sessions, while a premium title may have a smaller group that plays for many hours. Adding these together as a single “player count” can be misleading, because the two groups have little in common beyond the word “game”.
Revenue figures have a similar issue. A large market total typically bundles software, in-game purchases, subscriptions and sometimes hardware or advertising. Whether a given number includes each of these is rarely obvious from the headline.
Audiences who watch rather than play
Online gaming is also something people watch. Live streams and recorded videos of games draw viewers who may never play the title. Viewing figures are usually reported as peak viewers, average viewers or hours watched, and these behave differently. A single popular event can produce a short spike, while a steady channel may have a modest peak but a loyal audience across the year.
Viewership also tends to be affected by how platforms count. Embedded players, autoplay, and people leaving a stream running in a background tab can all raise totals without reflecting attention. How the audience is counted therefore matters as much as its apparent size. Organised competition has its own measurement questions, covered in the discussion of competitive esports, and the wider world of streaming video raises similar ones.
Surveys, panels and who gets asked
Independent estimates of gaming habits often come from surveys. Online survey panels are convenient, but they typically over-represent people who are comfortable online and already spend time on screens, which may overstate gaming. Telephone or in-person surveys may reach a different mix of people but are costlier and less common for this topic. The sampling frames behind a survey determine whom it can speak for, and a well-presented chart can hide a narrow sample.
Self-reports add further uncertainty. People tend to estimate their playing time imperfectly, and their answers may be rounded or shaped by how they want to be seen. Comparing self-reported behaviour with device or platform logs, where both exist, often shows differences, though logs have their own gaps, such as not covering every device someone uses.
Spotting a trend versus a snapshot
A rising line in a gaming chart can reflect real growth, but it can also reflect a change in what is counted, a new platform joining the dataset, a revised definition or a one-off event that drew people in. Periods of unusual circumstances, such as lockdowns, tend to raise play temporarily, and it is often unclear how much of that persists afterwards. Looking at whether a pattern holds across several independent sources, and over several years, is a more cautious approach than trusting one peak.
It also helps to ask what is actually changing: more people, the same people playing longer, or a growing share of existing players spending money. Each is a different story and can produce the same headline.
Questions worth asking of any gaming statistic
- What exactly was counted: accounts, devices, sessions or people?
- Over what time window, and is it a peak or an average?
- Who collected it, and what would they gain from a large number?
- Was it measured or self-reported, and who was asked?
- Does the comparison use the same definition across years and across countries?
If a figure cannot answer these, it may still be interesting, but it is better treated as an illustration than as a measurement. Online gaming is a large and varied activity, and it is probably safest to expect that the real picture is more fragmented than any single number suggests.