Most people who lose money betting do not lose it because their picks were bad. They lose it because their stakes were. A good run at random sizes followed by one big bet to win it all back is the most common story in the hobby. This guide covers the unglamorous part of betting that decides whether you are still enjoying it a year from now: how much money to set aside, how to size each bet, and how to stop a bad week becoming a bad month. It builds on the guide to margin and value, because staking only helps when the prices you take are sound.
What a bankroll is
Your bankroll is money set aside purely for betting. It is not your current account balance, not the money in your betting account this week, and not what you have left after the last bet. It is a fixed sum you decided on in advance, that you could lose entirely without it changing anything in your life.
That last part is the whole point. Betting with money you can afford to lose lets you make decisions on their merits. Betting with money you need makes every loss feel like an emergency, and emergencies produce the worst decisions in the sport: bigger stakes, longer shots and bets you would never have placed calmly.
A sensible first bankroll for someone learning is small, because the early months are tuition. Set an amount, deposit it, and treat topping it up as a decision to be made once a month at most, never in the middle of a losing evening.
Units: think in percentages, not pounds
Experienced bettors measure stakes in units rather than pounds. One unit is a fixed percentage of the bankroll, most commonly 1 percent, sometimes 2 for people who accept more variance. A £500 bankroll at 1 percent means a standard bet of £5.
Units do two things. They keep every stake proportionate to what you can afford, so a £5 bettor and a £500 bettor follow the same discipline. And they let you talk about results honestly: "up 12 units this month" means the same thing whatever your bankroll, while "up £60" tells you nothing without knowing how much was risked to get it.
Flat staking: the default plan
Flat staking means every bet gets the same stake, one unit, regardless of the odds or how confident you feel. It is the plan most professionals recommend to beginners and the plan many of them still use themselves.
It is simple. There is nothing to calculate and nothing to get wrong.
It removes emotion from sizing. Confidence is a feeling, not a probability, and beginners are poor at telling the two apart.
It makes your results readable. If you are losing with flat stakes, your selections are the problem, and you know exactly what to work on.
The one refinement worth considering is flat to win, where you size the stake so every bet would win the same amount. At odds of 3.00 you stake half a unit to win one unit; at 1.50 you stake two units to win one. This normalises the impact of each result, but it puts more money on favourites, which is exactly where beginners are most tempted to over-bet. Plain flat staking is the safer start.
Percentage staking
A variation is percentage staking, where each bet is a fixed percentage of the current bankroll rather than the starting one. After a winning run your stakes grow; after a losing run they shrink. The advantage is that you can never go completely bust, because the stake shrinks with the roll. The disadvantage is that recovering from a drawdown is slower, because you are betting smaller when you climb back. Both flat and percentage staking are respectable; percentage suits people who plan to bet for years, flat suits people who want the simplest possible rules.
Confidence tiers, used carefully
Some bettors use two or three stake sizes, half a unit for a speculative bet, one unit for a standard bet, two units for a strong one. This can work, but be honest about what "strong" means. If your two-unit bets do not win noticeably more often than your one-unit bets over a few hundred results, the tiers are just a way to lose money faster on the days you feel sure. Track them separately before trusting them.
The Kelly criterion, and why beginners should not use it
You will read about the Kelly criterion, a formula that gives the mathematically optimal stake for a bet where you know your edge. The stake is your edge divided by the odds minus one. For a bet at 2.20 where you believe the true probability is 50 percent, Kelly says stake about 8.3 percent of your bankroll.
The formula is correct and the problem is the input. It assumes you know your true edge precisely, and beginners do not. Overestimate the edge, which everyone does, and Kelly tells you to bet far too much; the formula punishes overconfidence brutally. Even experienced bettors who use it typically bet a quarter or a half of the Kelly stake. Know that it exists, and stick to flat units until you have several hundred tracked bets that prove your edge is real.
Staking plans to avoid
Chasing
Chasing means increasing your stake after a loss to win the money back. It is the most natural instinct in betting and the most destructive. The bet you place while chasing is the same bet with a bigger stake; the loss you just took has not made the next outcome any more likely. What chasing actually does is convert a normal losing run into a bankroll-ending one.
Martingale
The Martingale is chasing turned into a system: double your stake after every loss so that the first win recovers everything plus one unit. It sounds foolproof and it is not. Eight losses in a row at even money is a one-in-256 event, which happens to every regular bettor eventually, and it requires a stake 128 times your starting unit. Long before that the stake exceeds your bankroll, or the bookmaker's limit, and the system collapses at the worst possible moment.
Betting the balance
Treating whatever sits in your betting account as the stake for the next bet is not a plan, it is the absence of one. Winnings feel like free money and get bet larger; losses feel like a hole and get chased. Withdraw profits on a schedule and decide stakes from your bankroll, not your balance.
Understanding losing runs
Losing streaks are not a sign that something is wrong. They are arithmetic. Even a genuinely skilled bettor who wins 55 percent of even-money bets will see five losses in a row about once every 55 bets and seven in a row roughly once every 270. A recreational bettor without an edge sees them more often. At one unit a stake, a seven-bet losing run costs 7 percent of the bankroll, which is unpleasant and survivable. At 10 percent a bet it costs 70 percent and the hobby is over.
Stake per bet | Bets in a losing run | Bankroll lost |
|---|---|---|
1% | 7 | 7% |
2% | 7 | 14% |
5% | 7 | 35% |
10% | 7 | 70% |
This is the reason for the 1 to 2 percent rule. It is not about making money quickly. It is about being certain that a perfectly ordinary bad week cannot take you out of the game.
Keep a record
Write down every bet: the date, the selection, the odds, the stake, the bookmaker and the result. A spreadsheet is fine; several free apps do it on your phone. Records do three things no memory can. They tell you whether you are actually winning, because memory keeps the wins and misplaces the losses. They show which bet types and sports you are good at, so you can concentrate there. And they show you the size and frequency of your losing runs, which is the evidence you need to choose a stake size with a clear head rather than in the middle of one.
Set limits before you need them
Every licensed bookmaker offers deposit limits, which cap how much you can pay in over a day, week or month, and most offer loss limits and session reminders. Set them when you open the account, when you are calm and the bankroll decision is fresh. A limit set in advance is a plan; a limit set after a bad night is a reaction, and it usually comes too late.
Glossary
Bankroll: a fixed sum set aside for betting that you can afford to lose entirely.
Unit: a standard stake, usually 1 to 2 percent of the bankroll.
Flat staking: the same stake on every bet.
Percentage staking: a fixed percentage of the current bankroll on every bet.
Kelly criterion: a formula for the optimal stake given a known edge; unsuitable for beginners because the edge is not known.
Chasing: raising stakes after a loss to recover it.
Martingale: doubling the stake after every loss; fails when a normal losing run meets the bankroll or the bookmaker's limit.
Drawdown: the fall from a bankroll's peak to its subsequent low.
More in this series
This is part 4 of a six-part beginner's guide to sports betting.
