Key facts
- A betting bank is money set aside for betting and nothing else. Stakes are then sized as a share of that bank, not by how confident you feel.
- Stake size drives risk: at evens with a 5% margin against you, our calculation gives a 19.1% chance of halving the bank within 500 bets at 1% stakes, and 65.4% at 2%.
- A staking guide on the Betfair Hub suggests betting to collect 4% to 5% of your bank on win bets, which it says equals about 20% to 33% of a Kelly stake.
- Losing runs are normal: with a 30% strike rate there is an 83.5% chance of at least one run of 10 straight losers in 200 bets, in our calculation.
- Every licensed online wagering provider must offer deposit limits that cannot be raised on the spot, and send monthly activity statements.
What bankroll management is
Bankroll management is a set of rules for how much you bet: a fixed pot of money for betting, and a method for sizing each bet from that pot. It cannot make a losing punter a winner. What it does is control how fast money can go, so that one bad week does not decide everything.
The guide on the Betfair Hub defines the starting point well: "A betting bank is money that you put aside for the purpose of betting and nothing else." Rent, bills and savings are never part of it.
Step 1: set the bank
- Pick an amount you could lose in full without it changing your life. Treat it as already spent.
- Keep it separate. One betting account, funded once, makes it obvious when the bank is gone.
- Lock it in with a deposit limit. Under the National Consumer Protection Framework, every licensed online wagering provider must offer deposit limits, and they "cannot be increased on the spot". The TAB's limits run over 1, 7, 14 or 28 days; a lower limit applies straight away and a higher one only after 7 days.
- Decide what happens when it runs out. The answer should be "I stop", not "I top it up".
Our page on gambling deposit limits covers how to set them with each type of provider.
Step 2: size your stakes from the bank
The single biggest decision is the size of each bet as a share of the bank. To show why, we worked out the chance of losing half the bank within 500 bets, staking a flat amount on bets at $2.00:
| Flat stake (share of starting bank) | 5% margin against you | No edge either way | 3% edge in your favour |
|---|---|---|---|
| 1% | 19.1% | 2.5% | 0.5% |
| 2% | 65.4% | 26.4% | 11.0% |
| 5% | 90.9% | 65.5% | 45.5% |
| 10% | 96.3% | 82.3% | 68.4% |
Read the first column carefully, because it describes most punters: prices carry a margin, so the typical bettor is a little worse than even. At 1% stakes, about one in five punters halve their bank within 500 bets. At 5% stakes, nine in ten do. Even a punter with a genuine 3% edge has a 45.5% chance of halving the bank at 5% stakes. Small stakes do not change your edge; they change how long you get to find out what it is.
Four staking plans compared
| Plan | How it works | Strength | Weakness |
|---|---|---|---|
| Level stakes | The same amount on every bet, such as 1% of the starting bank | Simple, easy to track | Ignores the price: a $1.50 bet and a $15.00 bet risk the same |
| Percentage of bank | A fixed share of the current bank on each bet | Stakes shrink automatically in a losing run | Losses and wins of equal number still leave you behind |
| Bet to collect | Stake sized so a winner returns a set share of the bank | Risks less on longshots, more on favourites | Needs the bank recalculated regularly |
| Kelly (or a fraction of it) | Stake sized from your edge and the price | Grows the bank fastest if your edge is real | Badly overbets if your estimates are wrong |
Level stakes
Level staking is the easiest plan to stick to. The Betfair Hub guide gives an example of a $5,000 bank and a 1.3% level stake, $65 a bet, for a punter with a 30% strike rate. The weakness is that a level stake treats every price the same, so a run of longshot bets carries far more risk than a run of short-priced ones.
Percentage of bank, and the drag it creates
Staking a fixed share of your current bank means stakes fall as you lose, so you can never quite reach zero. The catch is that a win and a loss of the same size do not cancel out: a 10% stake that wins then loses leaves you on 0.9 x 1.1 = 99% of where you started.
Over 500 even-money bets that split exactly 250 wins and 250 losses, in our calculation, a $100 bank finishes on $90.48 at 2% stakes, $53.48 at 5% and $8.11 at 10%. Same results, very different outcome, purely from stake size.
Bet to collect
This plan sizes each stake so a winner returns a set amount. The Betfair Hub guide recommends betting to collect 4% to 5% of the bank on standard win bets. With a $10,000 bank and a $400 collect target:
| Price | Stake | Return if it wins |
|---|---|---|
| $2.00 | $200 | $400 |
| $3.50 | $114 | $400 |
| $4.00 | $100 | $400 |
| $8.00 | $50 | $400 |
| $20.00 | $20 | $400 |
The guide says this approach works out at roughly 20% to 33% of a full Kelly stake. It also suggests readjusting the target after every 10% to 20% move in the bank: a bank that rises from $2,000 to $2,200 lifts a $100 collect target to $110. Lay bets are sized the same way but by liability; our lay betting guide has that version.
Kelly staking
Kelly sizes a bet from the edge you believe you have. It only works if your probabilities are better than the market's, and it punishes overconfidence hard, which is why many punters use half or a quarter of it. The Kelly criterion calculator works out the stake, and our guide to value betting explains how to test whether your edge is real before you trust it.
Losing runs are part of the plan
A sound staking plan has to survive the runs of losers that every punter meets. In our calculation:
| Strike rate | Number of bets | Chance of at least one losing run of this length |
|---|---|---|
| 45% (fair price about $2.22) | 200 | 8 or more in a row: 53.0% |
| 45% | 200 | 5 or more in a row: 99.4% |
| 30% (fair price about $3.33) | 100 | 10 or more in a row: 58.0% |
| 30% | 200 | 10 or more in a row: 83.5% |
| 30% | 500 | 15 or more in a row: 50.9% |
A punter whose bets win 30% of the time should expect at least one run of ten straight losers more often than not, even over just 100 bets. With 10% stakes, that run alone takes about two thirds of a percentage-staked bank. With 1% stakes it is an annoyance.
Pundit tip: write your staking rule down before the first bet, including the stake, when it changes and the point at which you stop. A rule you invent after three losers is not a plan.
Never raise stakes to chase
Doubling up after a loss feels like control, but it is the opposite of bankroll management. It makes your biggest bets at the moment your bank is smallest. Our page on the martingale strategy shows how quickly that goes wrong.
Review the numbers each month
Licensed online wagering providers must send activity statements monthly under the National Consumer Protection Framework. Check yours against your own record of stakes, prices and results. If the bank is shrinking faster than the margin alone would explain, cut the stake size or take a break. If betting has stopped feeling like a hobby, our gambling help page lists free support, and Gambling Help Online is on 1800 858 858, 24 hours a day.
Frequently asked questions
What is bankroll management in betting?
It is the set of rules that decides how much you bet each time. You set aside a fixed betting bank, choose a staking plan, and size every bet from the bank rather than from a hunch, so a bad run cannot empty your account in a few days.
How much of my bankroll should I bet?
Smaller than most punters think. In our calculation, flat stakes of 1% of the bank at evens with a 5% margin against you have a 19.1% chance of halving the bank within 500 bets; at 5% stakes that rises to 90.9%. A staking guide on the Betfair Hub suggests betting to collect 4% to 5% of the bank on win bets.
What is the best staking plan?
There is no staking plan that turns losing bets into winning ones. Level stakes and percentage staking are simple and hard to abuse; Kelly staking only helps if your edge estimates are accurate, which is why many punters use a fraction of it. Plans that raise stakes after a loss are the ones to avoid.
Should I increase my stakes after a win?
Only by a rule set in advance. The Betfair Hub guide suggests readjusting stakes after every 10% to 20% move in the bank, up or down. Raising stakes because you feel on a hot streak, or to win back losses, is how a bank disappears.
Can I set a limit on how much I deposit?
Yes. Under the National Consumer Protection Framework every licensed online wagering provider must offer deposit limits, and they cannot be increased on the spot. The TAB, for example, offers limits over 1, 7, 14 or 28 days, with increases taking effect after 7 days.
Related reading
Betting AFL betting An evergreen guide to the AFL markets on an Australian betting slip: what each one asks, how the prices and the margin work, what happens when a game is drawn, and why you cannot bet in-play online. Read more 
Betting Futures betting A futures bet is placed now on a result that is weeks or months away: a premiership, a medal count, a big race. Longer prices come with a catch: most futures are All In. Read more 
Betting Tennis betting Tennis looks like the simplest bet in sport: two players, one winner. The catch is the retirement rule, which can void a bet on a player who was winning. Here are the markets, the maths and the rules. Read more 
Betting Head to head betting Head to head is the simplest bet in sport: pick who wins. The catch is in the draw rules, the margin on short favourites and the difference between two-way and three-way markets. Read more 
Betting Cricket betting Cricket has more ways to finish than any other sport on a betting slip: a win, a draw, a tie, a super over or a washout. Here is how each one settles, with the maths and the 2026-27 summer dates. Read more 
Betting First goal scorer Scorer markets pay big prices because only one player can score first. Here is how first and anytime markets differ, what the prices imply, and the rules for own goals, penalty tries and extra time. Read more Sources
Official pages we relied on. If anything here has changed, tell us via the contact page and we fix it.