Two bettors can pick the same winners and end up with very different bankrolls, because staking decides everything between winning picks and profit. A staking plan is simply a rule for how much to bet each time, and the right rule protects you during losing runs while letting winners compound.

Flat staking is the simplest plan: bet the same amount on every selection, often one or two percent of your starting bankroll. Its strength is predictability, since a losing run costs a known amount and emotions cannot inflate a bet after a win. Its weakness is that it ignores edge, betting the same on a marginal lean as on a strong value spot.

Percentage staking bets a fixed percentage of your current bankroll each time, so stakes shrink during losing runs and grow during winning ones. This built-in protection makes ruin far less likely than flat staking from a fixed starting point, and it is the plan most disciplined recreational bettors should start with, typically at one to two percent per bet.

The Kelly criterion is the mathematical approach: it sizes bets according to your estimated edge and the odds, betting more when the value is larger. In theory, full Kelly maximises long-term bankroll growth, but it is brutally volatile, and it assumes your edge estimate is accurate, which for most bettors it is not.

That accuracy problem is Kelly's fatal flaw in practice. Overestimate your edge by a few percent and Kelly tells you to bet far too much, turning a normal losing run into a bankroll crisis. This is why experienced bettors who use Kelly almost always use fractional Kelly, like quarter or half Kelly, which keeps most of the growth with far less volatility.

A worked example shows the difference. With a 1,000 dollar bankroll and a bet at even odds that you estimate wins 55 percent of the time, flat staking at 20 dollars risks little either way. Percentage staking at 2 percent risks 20 dollars now and less after losses. Full Kelly suggests 100 dollars, a tenth of the bankroll, on a single estimate that could easily be wrong.

Losing runs test every plan. At 2 percent percentage staking, ten straight losses still leave about 82 percent of your bankroll intact, and recovery needs only a modest winning run. At 10 percent stakes, the same run leaves roughly a third of the bankroll, and recovery requires nearly tripling what remains, a far harder task under pressure.

Confidence-based staking, where you bet more on games you feel good about, is the most common and most dangerous informal plan. Feelings after a win inflate stakes, feelings after a loss chase them, and the result is the largest bets landing on the worst emotional moments. A written rule beats a feeling every time.

Practical bettors often blend plans: percentage staking as the base, with a modest range, like one percent for standard bets and two percent for the strongest value spots, capped so no single bet exceeds that ceiling. This captures some of Kelly's logic, betting more on better value, without its volatility or its dependence on precise edge estimates.

Record-keeping makes any plan work. Log every bet with stake, odds, result, and the reason for the stake size, and review monthly. If your biggest stakes are losing more than your smallest ones, your confidence is misleading you, and the plan needs tightening, not loosening.

Choose by temperament as much as math. If you want simplicity and calm, flat or percentage staking wins. If you build models with genuine edge estimates and can tolerate swings, fractional Kelly is worth testing with small stakes first. If you cannot state your rule in one sentence, you do not have a plan yet.

Staking will not turn a losing picker into a winner, but it decides whether a winning picker keeps the profit. Pick a rule, write it down, follow it through losing runs, and let the plan, not the mood of the day, decide the size of every bet.