Bankroll management is the whole game
Spend time around experienced bettors and you’ll notice something odd: they talk about how much they bet almost as much as who they bet. That’s no accident. Picks get the attention, but staking is what determines whether you’re still betting next year. A bettor who wins 55% of their plays and sizes them sensibly grinds out a profit. A bettor with the same 55% who swings between 1% and 20% stakes can lose most of their money along the way, same picks, completely different outcome.
The first principle is separation. Your bankroll is not your checking account, your savings, or next month’s rent. It’s a dedicated pool of money set aside for betting, sized so that losing all of it wouldn’t change your life. When the pool shrinks, you don’t refill it from other money. When it grows, you don’t spend the growth on something else and call what’s left your bankroll. Keep it in its own place, a separate wallet, account, or at minimum a line in a spreadsheet, and treat its boundaries as rules, not suggestions.
Units: the 1-5 percent rule
A unit is your standard bet size, expressed as a percentage of your bankroll. Most bettors settle on one unit equal to 1% of the bankroll, with more aggressive players using 2% to 5% for their strongest plays. On a $2,000 bankroll that’s a $20 unit under the conservative approach, or up to $100 if you’re comfortable with more risk.
Why so small? Because of what variance does to big stakes. A bettor risking 20% of their bankroll per play who loses five in a row, entirely possible at any win rate, is down about 67%. A bettor risking 2% who loses the same five bets is down about 10%. Both streaks are equally likely; only one is survivable without a radical change of plan. The 1–5% range isn’t timidity. It’s what keeps a normal rough patch from becoming a bankroll-ending event.
Two practical notes. First, recalculate periodically: if your bankroll moves by 20% or more in either direction, reset your unit to match. Second, unit size and confidence can interact, many bettors use one unit as their base and size up to two or three units for their best plays. That’s fine as long as the maximum stays inside the 5% ceiling and you’re honest with yourself about what counts as a best play.
Flat staking vs percentage staking
Once you’ve picked a unit size, you have a choice to make about how it behaves over time.
Flat staking
With flat staking, every bet is the same amount, say, $20 on a $2,000 bankroll, regardless of how the bankroll moves. The advantages are simplicity and predictability: record-keeping is trivial, and your stake doesn’t shrink during a slump, so a losing streak doesn’t compound into ever-smaller bets. The trade-off is that a flat stake becomes a larger share of your bankroll as you lose and a smaller share as you win, the opposite of what risk management would suggest.
Percentage staking
With percentage staking, you bet a fixed percentage of your current bankroll, for example, 2%, recalculated after each session or each bet. After a losing streak your stakes automatically get smaller; after a winning run they compound upward. In concrete terms: start with $1,000 and lose ten straight bets at -110. Flat staking at $20 per bet leaves you with $800 and the same $20 stake. Percentage staking at 2% leaves you with about $817, and your next stake drops to roughly $16. The percentage approach is doing quiet risk management for you, at the cost of slightly more bookkeeping and the annoyance of smaller stakes exactly when you’re most eager to win it back.
Which should you use?
Either works. Flat staking suits bettors who value simplicity and discipline. Percentage staking suits bettors who want the math to adjust automatically. A common hybrid: use flat units, but re-baseline your unit whenever your bankroll moves by more than about 20%. Whatever you choose, write it down and follow it. The failure mode isn’t picking the “wrong” system, it’s switching systems mid-slump.
The Kelly criterion, in plain English
The Kelly criterion is a staking formula developed at Bell Labs in the 1950s that answers one question: given your edge, what share of your bankroll should you bet to maximize long-run growth? In its classic form it looks like this: f* = (bp − q) / b, where b is the net odds you’re getting, p is your estimated chance of winning, and q is your chance of losing. You don’t need to memorize it. What matters is the intuition: the bigger your edge and the better your odds, the more Kelly says to bet, and if you have no edge, Kelly says to bet nothing.
Run some real numbers through it. If you find a +150 underdog you honestly believe wins 45% of the time, full Kelly suggests staking about 8% of your bankroll. If you think a -110 favorite wins 55% of the time, an excellent edge, full Kelly says about 5.5%. Trim that estimate to a more realistic 53%, and Kelly drops to about 1.3%. Notice how violently the answer responds to small changes in your assumed win probability. That sensitivity is the whole story.
Why most bettors use fractional Kelly
Full Kelly has two problems in the real world. First, it assumes your probability estimates are correct. They aren’t: nobody’s are, at least not precisely. Feed Kelly a slightly inflated edge and it will happily oversize your bets. Second, even with perfect estimates, full Kelly produces stomach-churning swings; drawdowns of 30–50% are part of its normal behavior. Most practitioners therefore use fractional Kelly, half, quarter, or even a tenth of the suggested stake. Fractional Kelly gives up a little theoretical growth in exchange for dramatically less volatility and much more tolerance for estimation error. And in practice, fractional Kelly at half or quarter stakes lands you right back in the 1–5% unit range bettors have used forever. That convergence isn’t a coincidence; it’s different roads to the same sensible place.
Losing streaks are normal: the variance math
Here’s the fact that breaks more bankrolls than any other: losing streaks are not a sign that something is wrong. They’re arithmetic. Consider a good bettor: someone winning 55% of their bets. They lose 45% of the time. Four losses in a row has a probability of about 4% at any given starting point, and across 100 bets the chance of hitting at least one four-loss streak is about 90%. Across 200 bets it’s over 99%. A five- or six-loss streak isn’t rare either, just less frequent.
What this means practically: a bad weekend tells you almost nothing about your process, and a good weekend tells you almost as little. Short samples are dominated by variance, the random noise around your true edge. The gambler’s fallacy, the feeling that you’re “due” after a cold streak, is exactly backwards: past losses don’t make future wins more likely, they just make the bankroll math worse. The bettor who understands this keeps stakes steady and emotions level through swings in both directions. The bettor who doesn’t starts chasing, which is how a 10% drawdown becomes a 60% one. Rule of thumb: you need hundreds of bets, not dozens, before results start telling you anything reliable about your skill.
When to stop for the day
The last piece of bankroll management isn’t math: it’s knowing when to close the app. Set limits before you start playing, while you’re calm and objective:
- A stop-loss. Decide in advance that a certain loss ends the session: three units, or 5% of your bankroll, are common choices. When you hit it, stop. The games will still be there tomorrow.
- A time limit. Two hours, one game, or whatever fits your life. Betting past the point where you’re enjoying it is how losses accumulate for no good reason.
- Tilt checks. Tilt is the betting version of playing angry. Warning signs: raising stakes to get even, betting on games you don’t actually care about, hiding your activity from people close to you, or feeling irritable and distracted. Any of these means it’s time to stop, not to double down.
- No chasing. Chasing losses is the single most reliable way to turn a normal losing day into a disaster. The next bet is not a recovery mechanism.
Every licensed sportsbook offers practical tools for exactly this: deposit limits, session time-outs, cool-off periods, and self-exclusion. They take a minute to set, and they work because they bind future-you to present-you’s judgment. Our sportsbooks section covers what licensed operators offer and how to use these controls. And if betting has stopped feeling like entertainment, the National Problem Gambling Helpline is available 24/7 at 1-800-GAMBLER in the US.
The short version
Bankroll management isn’t the boring part of sports betting; it’s the part that decides whether you get to keep betting. Set a separate bankroll. Size your units between 1% and 5%. Pick flat or percentage staking and stick with it. Use Kelly as a sanity check rather than a command. Expect losing streaks, and don’t panic when they arrive. Walk away when you hit your limits. None of that makes you a winner by itself, but it’s what keeps a losing stretch from ending the story, and it’s what lets a real edge actually pay off over time.
