Gambling statistics: who gambles, how much, and how many are harmed

Most confusion about gambling numbers comes from mixing up different measures. Before comparing any two figures, check which of these each one reports.

Participation

Participation surveys ask whether people gambled at all within a period. Past-year figures are much higher than past-month figures, and some surveys count lottery tickets while others report them separately. A country where half the adults "gamble" may be a country where many people buy one lottery ticket a year.

Handle, revenue and yield

• Handle or turnover is the total amount wagered.
• Gross gaming revenue (GGR), called gross gaming yield in Britain, is what operators keep: stakes minus prizes. It equals what players lost in total.

The gap is large. If slot players put $100 million through machines returning 92%, handle is $100 million but revenue is about $8 million. If a sportsbook holds 8% of $1 billion in bets, revenue is about $80 million. Headlines that quote handle make an industry look many times bigger than one that quotes revenue. The house edge guide explains where that kept percentage comes from.

Prevalence of harm

Researchers measure harm with screening tools. The Problem Gambling Severity Index (PGSI) scores nine questions from 0 to 27; a score of 8 or more is usually classed as problem gambling, and lower scores as low or moderate risk. Clinicians use the DSM-5 definition of gambling disorder, which requires at least four of nine criteria within 12 months. The two tools do not produce identical numbers.

Method

Telephone, face-to-face and online surveys reach different people and get different answers. That alone can move a prevalence estimate several-fold, as the British example below shows. This guide is part of the Casino knowledge topic.

The best recent global estimate comes from a 2024 systematic review and meta-analysis in The Lancet Public Health, published alongside a Lancet commission on gambling. Pooling surveys from many countries, it estimated that about 46% of adults and about 18% of adolescents had gambled in the past year. Coverage was uneven: most data came from high-income countries, and many low- and middle-income countries had no national survey.

Selected figures

What the British numbers show

The Gambling Survey for Great Britain, run by the Gambling Commission, replaced older health-survey questions from 2023. The gap between 48% and 27% shows how much of everyday gambling is lottery draws. Once you take those out, a little over a quarter of adults had gambled on anything else in the past month.

Who gambles

Across most national surveys, men are more likely to gamble than women and more likely to use sports betting and online casino products. Lottery play is spread more evenly across ages and genders. Adolescent figures are notable because gambling is legally restricted to adults in most countries; much underage gambling happens informally or through products such as scratchcards and online games. The underage gambling guide covers those risks.

Market size is reported as revenue, which means player losses. Some reliable anchors:

United States

The American Gaming Association reported that US commercial gaming revenue reached a record of about $66.5 billion in 2023, covering commercial casinos, sports betting and online casino games in states where they are legal. That figure excludes tribal casinos, which report separately to the National Indian Gaming Commission and add tens of billions more. State lotteries are separate again; US lottery ticket sales have been reported at more than $100 billion a year in recent years, though sales are not the same as revenue because a large share is paid back as prizes.

Legal sports betting spread quickly after the US Supreme Court struck down the federal sports betting ban in Murphy v. NCAA in May 2018. States now regulate it individually, as the sports betting legal states guide explains.

Great Britain

The Gambling Commission's industry statistics have put annual gross gambling yield at around £15 billion in recent years, including the National Lottery, betting, casinos, bingo and online gambling.

Australia

Australia is often described as having the highest gambling losses per adult in the world. A 2023 federal parliamentary committee report on online gambling put total annual losses at roughly A$25 billion, with poker machines, known locally as pokies, the largest share.

Macau

Macau, the only part of China where casinos are legal, was the world's largest casino market by revenue before the pandemic, with 2019 gross gaming revenue of roughly US$36 billion. Revenue collapsed during travel restrictions and has been recovering since.

Why totals are rough

Global totals from consultancies differ because they include or exclude lotteries, unregulated offshore markets and online products differently. Treat any single worldwide number as an estimate, not a count.

Harm figures are the most sensitive and most disputed gambling statistics, largely because of the measurement issues described above.

United States

The National Council on Problem Gambling estimates that about 2.5 million US adults, around 1%, meet criteria for a severe gambling problem in a given year, and that another 5 to 8 million, around 2% to 3%, have mild or moderate problems. The DSM-5 text gives a past-year prevalence for gambling disorder of roughly 0.2% to 0.3% of the general population, with higher lifetime rates.

Great Britain

Older health surveys in England put problem gambling (PGSI 8 or more) at well under 1% of adults. The new online Gambling Survey for Great Britain produced a much higher figure, about 2.5% for 2023. The Commission cautioned that the two methods are not directly comparable: online self-completion may capture people who under-report in face-to-face interviews, and may also over-represent some groups. The truth is debated, which is a useful lesson in how method shapes numbers.

Worldwide

The 2024 Lancet review estimated that about 8.7% of adults showed some level of risky gambling, a broader category than clinical disorder. Rates were higher among people using online casino games and slots than among lottery players.

Beyond the gambler

Studies of gambling harm consistently find that partners, children, friends and colleagues are affected too, through debt, stress and broken trust. Estimates of how many people are affected by each problem gambler vary between studies, but most put it at several.

Behind every percentage are individuals. The gambling addiction guide covers signs and treatment, and PVPspinArena's responsible gambling page lists tools and support links.

Gambling generates a steady stream of numbers that circulate without sources. A few patterns to watch for:

Unsourced motivational claims

Lines such as "most gamblers quit just before a big win" appear in memes and forums with no study behind them. There is no way to measure it, and the logic fails anyway: each round is independent, so there is no "just before".

Survivorship

Winners talk and losers stay quiet. Streams, forums and adverts show the jackpots, never the thousands of sessions that funded them. Any sample drawn from people who chose to share results is biased toward wins.

Handle presented as size

A headline that "people bet $X billion" on an event is usually handle. The operators' revenue, and the players' net loss, is a small fraction of that.

Per-capita tricks

Dividing a national total by all residents, including children and non-gamblers, makes per-person figures look small. Dividing by gamblers only makes them look large. Both are legitimate if labelled; neither is if hidden.

A quick checklist

1. Who produced the number, and do they benefit from it?
2. What year does it describe?
3. Is it participation, handle, revenue or harm?
4. What screening tool or survey method was used?
5. Does it cover lotteries, online play and unregulated markets?

A statistic that cannot answer these questions is a talking point, not evidence.

National figures describe populations. For an individual player, the most useful number is the expected return on the bet in front of them.

PVPspinArena runs three player-vs-player games in USDC or ETH on Base. In Roulette, every bet on its 15-slot wheel returns 14/15, about 93.33%, on average, a 6.67% house edge. That means about $6.67 of expected loss for every $100 wagered, whatever colour you choose. A Coinflip is a fair 50/50, with the winner taking the pot minus any fee shown before entry. In Jackpot, your chance equals your share of the pot.

Checking the numbers yourself

Because results come from committed seeds, you can verify any settled round on the fairness page and build your own sample. Over 1,500 Roulette spins, Green should land about 1,500 ÷ 15 = 100 times. The standard deviation is the square root of 1,500 × (1/15) × (14/15), about 9.7, so anywhere from roughly 80 to 120 is normal. Results far outside that range over a large sample would be worth investigating; results inside it are the published odds working as designed.

The law of large numbers guide explains why a sample that size settles close to expectation, while a single evening does not. Gambling is 18+ only.

In the same cluster, see also sunk cost fallacy, loss aversion, and illusion of control.