RizonIs sports betting a good side hustle? What the numbers say
For most people, no. US sportsbooks keep about 10 cents of each dollar they take, and "97% lose" has no real source. What the data shows, and what it takes.
Short answer: For most people, no. US sportsbooks kept 10.16% of what customers put in during 2025, so the average customer is paying for the entertainment rather than earning from it. The few who do beat the line tend to get limited. That popular "97% of bettors lose" figure has no real source, but the data that does exist points the same way. If you think you have an edge, you can test it for free before risking anything.
What the data shows
You'll see "95% of bettors lose" everywhere. We couldn't find a study behind it, and sites repeating it admit as much. Sportsbooks don't publish the numbers. Here's the closest real data:
| Source | Who | Finding |
|---|---|---|
| American Gaming Association, 2025 | All US commercial sportsbooks | $16.96B kept on $166.94B taken in: 10.16% |
| Taylor, McCarthy and Wilbur, 2024 | 717,724 US card users over 4.75 years, all online gambling | Under 5% withdrew more than they spent |
| NatCen, UK, 2018 to 2019 | Seven operators' betting accounts | 23% won money over one year |
| Citizens JMP, on Juice Reel users, 2025 to 2026 | Sportsbook users who log their results | Median return -5%; the smallest users -29.3% |
| Bank of America Institute, 2026 | Card and bank flows | Customers recover under 75¢ of every dollar sent |
A year's winners and a lifetime's winners are different things. 23% can be up in one year by luck alone, and most of them won't stay up.
The sum for "$100 a day"
The vig decides this. At -110 you need 52.4% just to break even. Suppose you're genuinely very good, and win 55% over the long run. Then:
| Win rate (%) | Expected return per $1 risked (¢) | Risked per day for $100 a day ($) |
|---|---|---|
| 52.4 | 0 | Never |
| 54 | 3.1 | 3,200 |
| 55 | 5.0 | 2,000 |
| 56 | 6.9 | 1,450 |
That's in expectation, and the swings are large: even at 55%, a run of bad weeks is normal. And at those sizes sportsbooks notice. Regulators in Massachusetts heard that winners are likely to have their limits lowered, so the edge you worked for gets taken away. It's also not an hourly wage: you'd be doing research for free in the hope of an edge that may not exist.
Parlays make it worse. New Jersey sportsbooks held 19.2% on them in 2025, about four times what they held on everything else.
What about arbitrage and +EV?
Those are the serious attempts: taking prices that differ between books, or that a sharper book says are wrong. They rest on real maths, and they run straight into limits, since books cut the accounts doing it. They're also work, with capital tied up across many accounts. Treat claims of steady income from them with care, especially from people selling the software.
Better uses of the same instinct
If what you like is working out what a line should be, there's a more honest test. On an order book you can quote your own line and see who takes it, and nobody limits you for winning. Rizon's NFL books do this with practice money, with no vig and no sign-up. It pays nothing, because it's paper: it's for finding out whether you have an edge, not for earning. How sportsbooks make money shows what you'd be competing with.
If this has stopped being fun, or you're chasing losses, help is free: in the US, call 1-800-GAMBLER.
FAQ
Can you make a living from sports betting?
A very small number of professionals do, usually with large bankrolls, models and ways around limits. For nearly everyone else, the answer is no.
Can I make money from sports betting?
Over a short stretch, yes, by luck. Over the long run you need to beat the vig by more than chance, and the evidence says few do.
Do 97% of sports bettors lose?
Nobody has shown that figure. The closest real data: 23% of UK betting accounts won money in one year, and under 5% of a large US card panel withdrew more than they spent over nearly five years, across all online gambling.
How much do sportsbooks keep?
10.16% of what US customers put in during 2025, or about 4.5% on a balanced -110 market and much more on parlays. Here's why.
Is a prediction market better?
The fee can be smaller, and nobody limits you for winning, but retail users did worse in the one comparison we found.