The most overlooked skill in football betting is not handicapping, line shopping, or model building. It is deciding how much to bet. A bettor with a genuine 54% win rate on spreads will go broke if they stake 25% of their bankroll on every wager. A bettor with a 52% win rate will grind out a steady profit if they size their bets correctly and never deviate from the plan. Bankroll management does not generate edge — that comes from your analysis. It preserves the edge you have and gives it enough runway to compound into actual profit.
The concept is simple. You set aside a dedicated sum of money for betting — your bankroll. You then divide that bankroll into units and stake a fixed number of units per bet. The unit system creates a buffer between your emotional impulses and your capital, ensuring that no single loss can derail the entire operation. The details of how you size those units, and whether you adjust them over time, are where the meaningful differences between staking plans emerge.
Flat Betting: The Foundation of Staking Plans
Flat betting means wagering the same amount on every bet, regardless of how confident you feel. If your bankroll is $5,000 and your unit is $100 (2% of bankroll), you bet $100 on every single play. Whether it is a five-star lock or a marginal lean, the stake stays the same.
The appeal of flat betting is simplicity and emotional insulation. You never have to make a real-time decision about how much to risk, which eliminates a common source of error. Bettors who vary their stakes based on confidence tend to oversize their biggest bets — the ones they “love” — and those bets lose at the same rate as everything else. The perceived confidence does not translate into higher accuracy, but the larger stake translates into larger losses when the bet misses. Flat betting removes this trap entirely.
The downside is that flat betting does not optimize for the size of your edge. A bet where your model shows a 5% edge gets the same stake as a bet where the edge is 1%. In a perfect world, you would bet more on the larger edge and less on the smaller one. Flat betting sacrifices this optimization for consistency, which is a trade-off that favors the vast majority of recreational bettors whose edge estimates are not precise enough to guide variable sizing.
The standard recommendation for flat-bet unit sizing is 1-3% of your bankroll per bet. At 1%, a bankroll of $5,000 produces units of $50. At 3%, units are $150. The lower end of the range is appropriate for bettors who place many bets per week (10 or more) or who are still developing their handicapping skills. The higher end is suitable for bettors with proven track records and smaller bet volumes. Going above 5% per bet is rarely justified for any reason — the risk of ruin (total bankroll depletion) rises sharply beyond that threshold, even for bettors with a genuine edge.
Percentage-Based Staking
Percentage-based staking is a dynamic alternative to flat betting. Instead of betting a fixed dollar amount, you bet a fixed percentage of your current bankroll on each wager. If your bankroll grows, your bet size grows proportionally. If your bankroll shrinks, your bets shrink with it.
The advantage is automatic bankroll protection during losing streaks. If you start with $5,000 and bet 2% ($100), a ten-bet losing streak reduces your bankroll to roughly $4,000 and your next bet to $80. The declining bet size slows the rate of loss and makes it virtually impossible to go completely broke — you can always bet 2% of whatever remains. With flat betting, the same ten-bet streak costs $1,000 regardless of where the bankroll stands, and if the streak extends, the fixed bet size consumes an increasingly dangerous share of the remaining capital.
The disadvantage is that percentage staking makes recovery harder. After a losing streak, your bets are smaller, which means the winning bets that follow generate smaller returns. A bettor who lost $1,000 through flat betting recovers it in ten winning bets at $100 each. A bettor using percentage staking needs more wins to recover the same dollar amount because each winning bet is smaller than the losing bets that preceded it. The math favors percentage staking for survival but favors flat betting for recovery speed.
In practice, many experienced bettors use a hybrid approach: they recalculate their unit size periodically — weekly or monthly — based on the current bankroll, but keep the unit fixed within each recalculation period. This captures the bankroll-protection benefit of percentage staking while avoiding the constant recalculation that makes pure percentage staking cumbersome.
The Kelly Criterion: Optimizing for Growth
The Kelly Criterion is the most mathematically rigorous staking plan available, and it is also the most misunderstood. Developed by John Kelly at Bell Labs in 1956, the formula calculates the optimal bet size as a function of your edge and the odds offered. The goal is to maximize the long-term growth rate of your bankroll — not to maximize profit on any single bet, but to grow the total portfolio as efficiently as possible over time.
The formula for a standard -110 bet is: Kelly % = (bp – q) / b, where b is the decimal profit per dollar wagered (0.909 for a -110 bet), p is your estimated probability of winning, and q is the probability of losing (1 – p). If you believe you have a 55% chance of winning a -110 bet, the Kelly stake is (0.909 x 0.55 – 0.45) / 0.909 = 0.055, or about 5.5% of your bankroll.
That 5.5% figure is already aggressive by conventional standards, and it assumes your probability estimate is perfectly accurate — which it never is. This is the core problem with full Kelly staking: it is optimal only if your inputs are correct. If your estimated 55% probability is actually 52%, you are overbetting by a wide margin, and the excess bet size increases variance without increasing expected return. Overbetting even slightly beyond the Kelly fraction reduces long-term growth rate and increases the probability of significant drawdowns.
For this reason, virtually no serious bettor uses full Kelly. The standard practice is fractional Kelly — betting a fraction (typically one-quarter to one-half) of the Kelly-recommended amount. Half-Kelly on the previous example produces a 2.75% stake, which is well within the conventional 1-3% range and tolerates estimation error without catastrophic consequences. Quarter-Kelly is even more conservative and appropriate for bettors whose probability estimates carry significant uncertainty.
The Kelly Criterion is most useful as a conceptual framework rather than a literal calculation. It teaches the most important lesson in bankroll management: bet size should be proportional to edge. Larger edges justify larger stakes; smaller edges justify smaller ones. The exact formula matters less than the principle — and the principle is violated every time a bettor sizes a bet based on confidence, excitement, or a desire to recoup losses rather than on a quantified assessment of the edge at hand.
Comparing the Three Approaches
Each staking plan has a profile that suits a different type of bettor. Flat betting is best for beginners and recreational bettors who value simplicity and consistency. It requires no probability estimation, no formula, and no adjustment — just a fixed dollar amount per bet. The trade-off is that it does not optimize for edge, meaning you bet the same on a high-confidence play as on a marginal one.
Percentage-based staking is best for bettors who prioritize bankroll preservation and want automatic protection against extended losing streaks. It requires periodic recalculation but no probability estimation. The trade-off is slower recovery from drawdowns and slightly more administrative effort.
Fractional Kelly is best for sophisticated bettors who can produce calibrated probability estimates and want to optimize long-term bankroll growth. It requires the most analytical input but produces the most efficient capital allocation when the inputs are accurate. The trade-off is sensitivity to estimation error — garbage probabilities produce garbage sizing.
For most football bettors, the honest recommendation is flat betting at 1-3% of bankroll. It is not optimal in theory, but it is robust in practice, and robustness matters more than optimality in a world where probability estimates are uncertain, emotions are real, and the season is only 18 weeks long.
The Bankroll Is the Business
There is a useful thought experiment for anyone who struggles with bet sizing discipline. Imagine your bankroll is not money you set aside for entertainment. Imagine it is the working capital of a small business you own. Every bet is an inventory purchase. Every win is a sale. The vig is your cost of goods sold.
Would you invest 25% of your business’s working capital in a single product that might not sell? Would you double your inventory order after a bad quarter to try to make up the shortfall? Would you change your purchasing strategy every time a competitor had a good week? Of course not. You would size each purchase proportionally, track your margins, and adjust slowly based on data — not impulse.
The bankroll is the business. It is the infrastructure that allows your edge to express itself over hundreds of bets. Without it, the edge is theoretical — a number in a spreadsheet that never translates into dollars. Protecting it is not conservative. It is the bare minimum requirement for taking football betting seriously. Every dollar you preserve through disciplined sizing is a dollar that gets to participate in the next +EV opportunity, and the compounding effect of that reinvestment is the only mechanism through which small edges become meaningful profits. The bettors who survive are the ones who treat their bankroll with the same seriousness they bring to their handicapping. The ones who do not survive never had the chance to find out how good their picks actually were.
