Bankroll management for sports bettors comes down to two rules. First, bet only from a fixed pool of money you can afford to lose. Second, stake a small, consistent share of that pool on every bet, usually 1-3%.
Following those rules keeps a normal losing run from wiping you out. Bankroll management won't turn losing picks into winners. It keeps you betting long enough for a real edge, if you have one, to show up in your results.
Bankroll management for sports bettors in five steps
- Set the bankroll. Pick an amount you could lose in full without touching rent, bills or savings. Keep it in a separate account so it never mixes with everyday money.
- Choose a unit. A unit is a fixed share of the bankroll, commonly 1-3%. On a $2,000 bankroll, one unit is $20-60.
- Stake in units, not dollars. A standard bet is 1 unit. Many unit bettors go to 2-3 units on their strongest spots and never go above 5 units on a single bet.
- Log every bet. Record the date, market, odds, stake, result and sportsbook. Without a record, you can't tell skill from luck.
- Set review points in advance. Decide when you will stop and reassess, for example after losing 25% of the starting bankroll. Also decide when you will recalculate the unit.
Unit sizes at 1%, 2% and 3% for common bankrolls
1% of a bankroll is one hundredth of it. That works out to $1 on a $100 bankroll and $10 on $1,000.
| Bankroll | 1% unit | 2% unit | 3% unit |
|---|---|---|---|
| $100 | $1 | $2 | $3 |
| $500 | $5 | $10 | $15 |
| $1,000 | $10 | $20 | $30 |
| $5,000 | $50 | $100 | $150 |
Unit = bankroll × stake percentage.
A $1 unit feels small, and that is deliberate. At flat 1% stakes, it takes 100 straight losses to reach zero.
Flat, percentage or Kelly: choosing a staking plan
Bankroll guides disagree on which staking plan is best. Each plan balances simplicity against how quickly stakes react to wins and losses. For most bettors, we suggest starting with flat units and moving to another plan only once the bet log has real volume.
| Plan | How the stake is set | Strength | Weakness |
|---|---|---|---|
| Flat units | Same amount every bet, based on the starting bankroll | Easy to follow and track | Stake stays the same during a losing run |
| Percentage | Fixed % of the bankroll as it stands today | Stakes shrink after losses and grow after wins | Slower to recover after a bad run |
| Confidence tiers | 1-3 units, depending on how strongly you rate the bet | More money on your best spots | Confidence is not a measured edge |
| Fractional Kelly | A quarter to half of the Kelly figure | Stake reflects both the price and your edge | Only as good as your win-probability estimate |
Kelly Criterion formula and a worked example
John L. Kelly Jr. published the Kelly Criterion in the Bell System Technical Journal in 1956. It gives the share of your bankroll that produces the fastest long-run growth, as long as your probability estimate is correct.
Kelly % = (bp − q) / b
- b = decimal odds − 1 (profit per $1 staked)
- p = your estimated chance of winning
- q = 1 − p
Worked example: you think a bet priced at -110 (decimal 1.909) wins 55% of the time.
- b = 0.909, p = 0.55, q = 0.45
- Kelly % = (0.909 × 0.55 − 0.45) / 0.909 = 5.5% of bankroll
The weak point is the estimate. If your true win rate is 53% and not 55%, Kelly says to stake 1.3%, so a 5.5% stake is more than four times too big. Fractional Kelly leaves room for being wrong about p.
What is a good ROI for a sports bettor?
At -110 you risk $110 to win $100. That means you need to win 52.4% of your bets just to break even (110 ÷ 210). ROI is profit as a share of total money staked. Each percentage point of win rate above break-even adds roughly 2 points of ROI.
| Win rate at -110 | ROI per bet | Full Kelly stake | Quarter Kelly stake |
|---|---|---|---|
| 52.4% | 0.0% | 0.0% | 0.0% |
| 53% | 1.2% | 1.3% | 0.3% |
| 54% | 3.1% | 3.4% | 0.85% |
| 55% | 5.0% | 5.5% | 1.4% |
| 57% | 8.8% | 9.7% | 2.4% |
| 60% | 14.5% | 16.0% | 4.0% |
Our calculations: ROI = p × 0.909 − (1 − p); Kelly stake = ROI ÷ 0.909, as a share of bankroll.
A sustained ROI of 3-5% on sides and totals already requires a 54-55% win rate. A tipster claiming 20% long-term ROI on those markets is claiming to win nearly 63% of bets.
The table also shows why 1-3% units make a sensible default. At 54-55% win rates, quarter-to-half Kelly works out to between about 0.85% and 2.75% per bet.
Price matters as much as picks. At -105, break-even drops to 51.2%, so the same record earns more.
Results also need volume before they mean much. A bettor with no edge at all goes 58-42 or better over 100 bets roughly 15% of the time.
How long losing runs get, even at a 55% win rate
Strong bettors still lose several bets in a row, and it happens often. Over 500 bets at a 55% win rate, the longest losing run you should expect is about seven straight. How much damage that run does depends on your stake size.
| Stake per bet (% of current bankroll) | Bankroll left after 7 straight losses | Gain needed to get back |
|---|---|---|
| 1% | 93.2% | +7.3% |
| 2% | 86.8% | +15.2% |
| 3% | 80.8% | +23.8% |
| 5.5% (full Kelly, 55% example) | 67.3% | +48.6% |
| 10% | 47.8% | +109.1% |
Our calculations, assuming percentage staking. The longest-run estimate uses the standard approximation for runs in a series of independent bets.
At 1-2% stakes, a seven-bet slide is an annoying stretch. At 10% stakes it halves the bankroll, and you then need to more than double what's left just to get back to where you started.
Six habits that drain a betting bankroll
- Chasing losses. Raising stakes to win back what you lost turns a normal losing run into the 10% row above.
- Betting on tilt. Bets placed out of frustration after a bad beat skip your own process. Step away until you would place the same bet on a calm day.
- Raising units after a hot streak. A good month is well within normal luck for a break-even bettor. Resize on a schedule, not on a feeling.
- Making parlays your main strategy. The sportsbook's margin compounds with every leg. Suppose each leg is a true 50/50 priced at -110. The book then keeps about 4.5% of a single bet, 8.9% of a two-leg parlay and 13.0% of a three-leg parlay.
- Stacking correlated bets. A side, a total and three player props on the same game can put 5 units on one result. Count them as a single exposure.
- Not tracking. Without a log, your ROI is a guess.
How bankroll rules carry over to prediction markets
The same principles apply to trading event contracts, with a few adjustments:
- Position size. Cap the capital you commit to any one market, the same way you cap units per bet.
- Unrelated markets. Five positions tied to one election or one game count as one position.
- Liquidity. Keep positions smaller in thin markets, because exiting early can mean selling well below fair value.
- Open positions. Track the current value of what you hold, not only settled profits.
- Fees. On platforms that charge trading fees, the fees raise your break-even the way vig does at a sportsbook. See what a prediction market trade actually costs.
When to resize units and withdraw profits
Recalculate your unit on a fixed schedule, not after every result. Two common approaches:
- Calendar: review your win rate, ROI and average odds each month, then reset the unit each quarter to 1-3% of the current bankroll.
- Threshold: reset the unit whenever the bankroll moves by a set amount, for example 20% up or down from the last reset.
Resizing down is the step bettors most often skip. Take a $1,000 bankroll with $30 units that falls to $700. At the same 3%, the unit should now be $21. Keeping $30 stakes means betting over 4% of what's left.
Withdrawing part of your profit at each reset locks in some of a winning run. It also keeps the pool at a size you chose on purpose.
Better prices lower your break-even on every bet: compare odds across sportsbooks.