staking plans for bettors: bank management that keeps you in the game

what a staking plan is

An edge in betting is a long-run property: it pays out over hundreds of bets, not on any single one. But the bank is spent one bet at a time, and the short run is noisy enough to bury a genuine edge for weeks. A staking plan is the rule that decides how much of the bank each bet risks, and it does three jobs. It keeps the bank alive through the bad runs that every record will contain. It scales the stake with the size of the edge, so bigger edges risk more and guesses risk less. And it makes the decision boring, because a stake size chosen in the heat of a losing evening is not a plan, it is a mood. Every plan starts from the same raw material, the price converted into a probability, which is the subject of the implied probability guide; staking is what you do with the number that conversion produces.

flat staking: a fixed amount, every time

Flat staking risks the same amount on every bet, however the results have gone. The amount can be a fixed sum, £20 a bet, or a fixed fraction of the original bank, 2% of the £1,000 you started with, which is the same thing: £20 every time. The arithmetic is instant, the record is easy to audit, and the risk per bet is known to the penny in advance. One hundred flat £20 bets can lose you at most £2,000, and fifty can lose you at most £1,000, the whole bank, no faster. The weakness is that the plan does not adapt. After ten losses the bank is £800, and the same flat £20 is now a 2.5% stake of what is left, so a flat plan lets the effective risk creep up exactly when the bank is weakest.

percentage staking: a fraction of the current bank

Percentage staking re-prices the stake on every bet: 2% of whatever the bank is right now. After the same ten losses it is 2% of £817, not 2% of £1,000, so the stake shrinks as the bank shrinks and grows as it grows. This is the compounding family of plans, and its best-known member is the kelly criterion guide, which sizes the bet as a fraction of the bank that maximises long-run growth for a given edge. Percentage plans feel natural once you think in fractions, but they demand one discipline: a true, current valuation of the bank, because the whole point is that the stake follows the money, not the memory of it.

worked example: a ten-bet losing streak

Set a £1,000 bank and give it the worst stretch a normal betting life contains: ten straight losing bets. At even-money prices with a genuine 50% chance on each bet, a run of ten losers is a 1-in-1,024 sequence, and over a thousand bets you should expect to see about one of them. Not a freak event: a scheduled one, and the plan decides whether the bank is still standing when it arrives.

stake level	flat, on the original bank	percentage, of the current bank
2% of the bank	£800, down 20%	£817, down 18%
5% of the bank	£500, down 50%	£599, down 40%
10% of the bank	£0, bust	£349, down 65%

Read the bottom row first. At 10% stakes the flat plan is finished: ten scheduled losers and the betting career is over. The percentage version survives but barely: £349 left, and getting back to £1,000 needs a gain of about 186%, a near-perfect season. The 2% current-bank plan is down 18% and needs a 22% gain to recover, which is the difference between a bad month and a career event. That is the whole argument for small stakes in one table: the cost of the same losing streak is not proportional to the stake, because recovery compounds. Halving the bank doubles the work to rebuild it, and the deeper the hole, the faster the rebuild arithmetic outruns the stake numbers.

the martingale: why loss-chasing kills banks

There is one plan this site will say plainly to refuse: the martingale, which doubles the stake after every loss so that the next win recovers everything plus one unit. At even money, starting at £10, the sequence is 10, 20, 40, 80, 160, 320: £630 committed to win £10. Six straight losers at even money is only a 1-in-64 event, so the plan wins small, often, and feels bulletproof for months, then meets the one losing run that exceeds the bank. A ten-loss chase would need £10,230 committed to win £10, and the previous table showed a ten-loss run is an expected event over a thousand bets. The maths is not on the plan's side; it is a short-run illusion sold against a long-run certainty. Loss-chasing in any form, doubling, raising after losses, topping the bank up mid-run, is the classic bank killer, because it does the opposite of every survival rule above: it makes the stake largest when the bank is smallest and the pressure is highest. Flat and percentage plans cap the damage at a number agreed in advance; the martingale uncaps it at exactly the wrong moment.

the kelly fraction: sizing the edge, not the hope

The percentage family has a theory behind it. Take a price of 2.50 and rate the true chance at 45%. The edge is 12.5p per £1 staked, and the Kelly fraction, edge divided by odds minus one, is 8.33% of the bank: full Kelly. Half Kelly is 4.17%, quarter Kelly is 2.08%. Notice where quarter Kelly lands: right at the 2% that the flat plan chose by feel. Full Kelly grows the bank fastest in expectation and swings it hardest in practice, with drawdowns of half the bank or more along the way, which is why most practical bettors run half or quarter Kelly. The staking calculator does all three fractions from bankroll, odds and probability, plus points staking. The deeper point is that a stake is not a second opinion. Once you have written down your probability, the stake is arithmetic, the subject of the kelly criterion guide; the only judgement left is how much of the theoretical maximum your stomach can hold.

variance: the edge is a long-run number

Staking matters because variance is what betting actually is, moment to moment. Bet £20 at 2.50 on a genuine 45% chance and the expected return is £2.50 per bet, but the standard deviation of a single bet is about £24.87, more than the stake. A single result says almost nothing about the edge. The run says more, slowly: after twenty such bets the chance of being in profit is only about 59%, after a hundred it is about 82%, after five hundred about 99%. That is the honest shape of an edge: a fifth of the time, after a hundred genuinely good bets, you are still losing money, not because the edge is false but because the variance has not finished with you. The stake is the only dial on that picture. The same edge at 10% stakes can halve the bank before it pays; at 2% it costs the bank a bad month and no more. The odds converter turns your prices into the probabilities this arithmetic needs.

staking as risk-of-ruin control

Put it together and staking is best understood as risk-of-ruin control. The question a plan answers is not how much you can win; it is how likely the bank is to hit zero before the edge pays. At 2% flat stakes, busting means fifty straight losing bets, at even money an event on the scale of one in a quadrillion; at 10% flat it means ten straight losers, an event that shows up about once per thousand bets. One of those is a scheduled appointment; the other is effectively never, and the plan chooses which you are betting against. Staking is the other half of this site's method. The value betting guide finds the edge; the plan keeps you alive to collect it. The bookmaker margins guide strips the one cost in the price you can control, and the stake caps the one risk in your own behaviour you can control. The dutching guide splits stakes across a shortlist, and the arbitrage betting guide is the rare case where the stakes are fixed by the market itself. Every one of them is a plan; none of them is a reason to bet more than the plan says.

the staking checklist

Before the first bet of any plan: fix the bank at a number you can afford to lose completely, and never top it up mid-run. Choose the plan in writing before the results start, flat 1-2% of the original bank, or 1-2% of the current bank, or quarter Kelly from the staking calculator, and write the stake down before you write the bet down. Never raise a stake after a loss; a raise after a loss is the martingale in disguise, whatever you call it. Re-baseline the bank weekly or monthly, never bet by bet. Expect streaks: a ten-loss run is a scheduled event over a thousand bets, not a signal that the plan is broken. If the bank falls below two-thirds of its starting level, halve the stakes until it climbs back; the recovery maths below that line is brutal. And remember an each-way bet is two stakes in one, win part and place part, sized together as the each-way betting guide shows. A plan that survives these rules is not exciting; that is the point.

the honest endnote

No staking plan turns a losing opinion into a winning one. The plan is the seatbelt, not the engine: the engine is the edge, the margin shopping keeps it, and the plan makes sure you are still in the car when it pays. Flat 1-2% of a bank you can afford to lose, never chase, never double, and the worst the game can do to you is a boring month instead of a busted bank. Bet within your means, treat the bankroll as money you can afford to lose, and if betting stops being enjoyable, stop. 18+ | BeGambleAware.