How to bet on tennis: bankroll, value betting and risk

Most people who bet on tennis lose money. That isn’t a pessimistic take, it’s the mechanical result of the bookmaker’s margin: every bet carries a small commission, and you have to be better than the market just to pay it back before you win anything.

We know this better than most because we tried. Pronocast computes a win probability for hundreds of matches every day, and we tested what happens if you bet systematically whenever our model disagrees with the odds. The result is further down, and it isn’t flattering. It does, however, explain better than any theory what actually matters when you bet: how much you stake, what value really means, and how to survive the losing runs.

This guide won’t give you a way to win for sure, because there isn’t one. It will give you ways to lose less, lose more slowly, and tell whether you have a real edge or just had a good run.

Before your first bet: what the odds are asking of you

Decimal odds read as a probability: 1 divided by the odds. Odds of 1.80 imply a 55.6% chance of winning, odds of 2.20 imply 45.5%. We walk through that calculation, and how to strip out the bookmaker’s margin, in our piece on tennis odds and implied probability.

Flip that number around and you get your break-even point. If you usually bet at around 1.80, you need to win more than 55.6% of your bets just to stay level. At 1.50 you need 66.7%. Plenty of bettors who back “safe” favourites have a hit rate that looks excellent to them, 62 or 63%, and still lose, because their odds demanded more.

Your bankroll: money you are prepared to lose

Your bankroll is the money you set aside for betting, kept apart from everything else. The definition that matters is this one: an amount you could lose entirely without it changing anything about your month. Not rent money, not savings you’ll “put back later”. If losing it would hurt, it’s too big.

The second rule is harder to keep: you don’t top it up after a losing run. Adding money to “win it back” is exactly how a small loss turns into a big one. If the bankroll hits zero, the experiment is over, and that is useful information in itself.

Flat staking, the most underrated method

Flat staking means always betting the same amount, a small fraction of your starting bankroll: usually 1 to 2%. With £500, that’s stakes of £5 to £10. It’s small, and that’s the point. This method has one advantage the others don’t: it doesn’t rely on any estimate of yours. You can’t get the stake wrong, only the pick.

Proportional staking

A variant: stake a fixed percentage of your current bankroll rather than the starting one. After a losing run your stakes shrink on their own, which slows the fall. After a winning run they grow. It protects you better than flat staking in bad spells, at the cost of recalculating before every bet.

The Kelly criterion, and why to be wary of it

The Kelly criterion, published by John Kelly in 1956 in the Bell System Technical Journal, gives the fraction of your bankroll that maximises its long-run growth: f = (p × o − 1) ÷ (o − 1), where p is your probability of winning and o the decimal odds.

On paper it’s the optimal stake. In practice one example shows the problem. You think a player has a 60% chance of winning and they’re priced at 1.90. Kelly tells you to stake 15.6% of your bankroll. Now suppose you were a little optimistic and the true probability is 55%. With the right number, Kelly would have said 5%. A five-point error in your probability tripled your stake.

Nobody knows their probability to within five points. That’s why people who use Kelly bet a fraction of it, a quarter or a half, and why for most bettors flat staking remains the safer choice. Kelly is a good tool for someone whose probabilities have proven themselves over thousands of bets. For everyone else, it speeds up the losses.

Losing runs happen, even when you bet well

Take a bettor who wins 55% of their bets, which is already very good. We simulated a 500-bet season for that profile 20,000 times. In 90% of cases they hit at least one run of 6 straight losses. In more than one case in three, a run of 8. These aren’t accidents, just the normal variance of something where you lose 45% of the time.

With 2% stakes, eight losses cost 16% of your bankroll: unpleasant, but you carry on. With 10% stakes, the same run wipes out 80%. That’s the real reason to stake small: not caution for its own sake, but the fact that the bad run will come and you need to still be there afterwards.

Value betting, explained without the sales pitch

A bet has value when the odds pay more than the true probability justifies. The maths is simple: probability × odds − 1. If you think a player has a 50% chance and they’re priced at 2.20, the expected value is 0.50 × 2.20 − 1 = +10% per unit staked. Over the long run, only positive expected value bets make money.

It all hangs on one word: true. Value only exists if your probability is more accurate than the market’s. And the market is the combined money of tens of thousands of bettors, including professionals who hammer mispriced odds until they’re no longer worth attacking. Spotting a gap between your estimate and the odds is easy. Knowing whether it’s you or the market that’s right is the whole problem.

What we measured ourselves

We put the question to our own model. Across 33,098 matches played between January 2025 and August 2026, we simulated a bet every time the model rated a player at least 5 probability points above what the market’s odds implied. That triggered 17,740 bets, of which 47.6% won.

The result: a return of −9.01%. Negative on the ATP (−6.7%), on the WTA (−8.8%), and on both men’s and women’s Challengers (around −9.9%). We were hoping to find a niche in the Challengers, which get less attention, where the market might be less sharp. It wasn’t there.

The explanation makes sense once you spell it out. The market was more accurate than our model at the time. When two estimates disagree and one is more reliable, the gap usually points to the mistake of the less reliable one. Betting on those gaps meant betting, match after match, against the side that was right more often. Since then the model has caught up with the market (69.0% of predictions correct against 69.2%, the details are in our methodology), but matching the market still gives you no edge to exploit against it.

We’re telling you this because it’s the most common mistake among people discovering value betting: believing a gap between your view and the odds is an opportunity. Most of the time it’s a warning.

Where you can genuinely gain a few points

There’s one lever that doesn’t require outsmarting the market: taking the best available price. Between 1.80 and 1.87 on the same player, your winnings are 3.9% higher for exactly the same prediction. Repeated over hundreds of bets, that gap weighs more than most “analysis”. It’s also the measure serious bettors track: do they get, on average, better odds than the final price just before the match? If so, they’re taking value before it disappears. If not, no hit rate will make up for it in the long run.

In France, where Pronocast is based, operators licensed by the Autorité nationale des jeux have a payout rate capped by law, at 85% on average for sports betting. Odds there are structurally lower than elsewhere, which raises the bar for being a winning bettor even further.

What’s specific to tennis

Retirements

Players retiring mid-match is common in tennis, and every bookmaker has its own rule: some void the bet and refund your stake, others settle it as soon as one full set has been played. The same retirement can be a lost bet with one and a refund with another. Read your bookmaker’s rule before betting, especially on a player coming back from injury.

Smaller tournaments

Challengers and ITF events look like ideal ground: less information, little-known players, odds you’d imagine are rougher. Our backtest says otherwise, with returns as negative as anywhere else. Add margins that are often higher and results that are more erratic. The lack of information works against you as much as against the bookmaker.

In-play betting

Betting during the match is the fastest way to lose your discipline. Odds move on every point, the margin is usually higher than before the match, and the decision is made in seconds, often right after watching the player you’re backing drop serve. If you bet in-play, set yourself a limit before the first point.

Accumulators

An accumulator multiplies the odds, and it multiplies the margins too. With a 5% margin per match, a single bet costs you about 4.8% in expected value. A four-leg accumulator costs you 17.7%, a six-leg one 25.4%. That’s why bookmakers push them so hard: they’re their most profitable product.

Managing risk day to day

• Log every bet: date, match, odds taken, stake, result, and the reason for the bet in one sentence. Without that record you’ll remember the wins and forget the losses, like everyone does.
• Set limits on your account. In France, licensed operators must make you set a deposit and stake limit when you open your account, and most regulated markets offer the same tools. Set them low: it’s easier to raise them with a cool head than to respect them in the heat of the moment.
• Don’t bet to win it back. A bet placed because you just lost is almost always a bad bet.
• Don’t judge anything on 100 bets. A bettor who wins 55% of their bets at average odds of 1.85 has an expected return of +1.75%. Over 100 bets, their measured return could land anywhere between roughly −16% and +20%. Over 1,000 bets the range is still about 6 points either way. In other words, a good month proves nothing, and neither does a bad one.

What Pronocast can do for you, and what it doesn’t

Pronocast isn’t a tipster service. For every match we publish a calibrated probability, computed before the match and frozen, then the result, on pages anyone can check. That’s useful for two things: getting a second, numbers-based opinion on a match, and checking for yourself what that second opinion is worth. Today’s picks are on the tennis predictions page, tour by tour, and where the probabilities come from is explained in our piece on Elo in tennis.

What we won’t do is promise that following our predictions will make you money. Our own numbers say otherwise, and we’d rather tell you here than let you find out from your bank statement.

Gambling involves risks: debt, isolation, addiction. Betting is for adults only. In France, help is available on 09 74 75 13 13 (Joueurs Info Service, free of charge). In the UK, call the National Gambling Helpline on 0808 8020 133. Elsewhere, your national gambling regulator lists a free helpline.