Risk of Ruin: How Much to Stake Per Bet

What Risk of Ruin Actually Means

Every punter asks "will this bet win." Almost nobody asks the question that actually decides whether they're still betting in six months: what happens to my bank if it doesn't.

That's risk of ruin. Not the chance of losing one bet, but the chance that your staking, run over enough bets, eventually empties the bank completely. Two different questions. A bet can be a good bet and still be sized in a way that buries you over a bad month.

Here's the part most staking advice skips: stake size feels the same in the moment whether it's 2% or 20%. You back the same team, you like the price the same amount, the confidence reads identical. The gap only opens up over a run of results, when a losing streak that was always going to happen at some point actually turns up. A bettor with a real edge can still go broke if the stakes are too big for the bank to absorb a bad run. That's not bad luck. That's bad sizing.

Risk of ruin puts a number on it. Feed in your bankroll, your stake size, your win rate, and it spits out the odds of hitting zero. Same picks, same odds, same edge, stake 1% and you're nowhere near that number. Stake 10% and you might be flirting with it every week.

Why "How Much Should I Bet" Is the Wrong First Question

Most bettors start with the pick, then figure out the stake almost as an afterthought. Fancy a team at 2.10, feeling good about it, so it's a tenner. Feeling great about it, so it's twenty. That's the entire process for a lot of punters, and it's backwards.

The stake shouldn't come from how confident you feel. It should come from what your bank can survive.

Confidence is a terrible sizing tool because it's the same regardless of the actual math. You can be just as confident about a bet with a 5% edge as one with a 15% edge, the feeling doesn't scale with the number, only the outcome does, eventually, after the damage is done. Sizing off confidence means your biggest stakes go on the bets that felt best, not the bets that were mathematically best. Those aren't always the same bet.

Flip the order round. Start with the bank. Decide, before a single price is checked, what fraction of it you're willing to put on any one selection. That number stays fixed regardless of how strongly you fancy the game. Once that's locked in, the pick-picking can happen separately, on its own merits, without a stake size trying to express how sure you are.

This is the entire logic behind a staking plan, and it's why "how much should I bet on this" is the wrong question to be asking match by match. The right question gets asked once, covers every bet going forward, and it's this: given my bank, what stake keeps my risk of ruin somewhere sane.

The Risk of Ruin Formula

Strip it down to its simplest form and the risk of ruin formula looks like this:

Risk of Ruin = (q ÷ p) ^ (B ÷ S)

Where p is your win rate, q is 1 minus your win rate (your loss rate), B is your bankroll expressed in stake units, and S is your stake size. That exponent, bankroll divided by stake, is doing almost all the work here. It's counting how many losing bets in a row it would take to wipe you out, and the formula is asking how likely that run actually is.

This version assumes flat, even-money stakes. Real bookmaker prices rarely land exactly on evens, so it's a simplified model, not the full picture, worth saying plainly rather than dressing it up as more precise than it is. It's still the right starting point because it shows the mechanic clearly before more variables get added.

Run the numbers. A £1,000 bank, staking £20 a bet, that's 2% a time, or 50 units. A 55% win rate. Plug it in: q/p works out to roughly 0.818, raised to the power of 50. That comes out to something like 0.004%. Effectively zero. At 2% a bet, with a real edge behind you, going bust just isn't a live concern.

Now take the same bank, same win rate, but stake £200 a bet instead. Still £1,000, still 55%, but now it's 5 units instead of 50. Same q/p, this time raised to the power of 5. That comes out around 37%. More than a one-in-three shot at going to zero, on the exact same picks, at the exact same price, with the exact same edge. The only thing that moved was the stake.

Stake Size vs Risk of Ruin

Same bank, same edge, only the stake size moves. £1,000 bankroll, 55% win rate, even-money bets throughout:

Read that table properly and it stops being an abstract percentage and starts being a decision. At 2%, risk of ruin is a rounding error. At 5%, it's real but manageable, under 2 in 100. By 10%, you're at roughly 1 in 7. By 20%, you're closer to a coin flip than most bettors would ever knowingly accept if you asked them straight out "do you want a 1-in-3 chance of losing your entire bank."

Nobody sets out to stake 20% a bet. It creeps there, a chase after a bad week, a "lock" that deserves double the usual stake, a bonus that gets treated as house money and staked accordingly. The table is the same regardless of how the stake got that big. Risk of ruin doesn't care about the story behind the number, only the number.

Staking Plans: Flat, Percentage, and Where a Masaniello Cycle Differs

Flat staking is the blunt instrument. Same £ amount every bet, win or lose, bank size doesn't move the number at all. It's the easiest plan to run and the easiest to understand, which is exactly why it's the default most bettors fall into without ever deciding to.

Percentage staking fixes the flaw flat staking has by design: it can't tell the difference between a healthy bank and a bank that's already bleeding. Percentage staking stakes a fixed share of the current bank, not the starting bank, so the £ figure shrinks automatically during a bad run and grows automatically during a good one. That's the mechanism doing the risk-of-ruin work in the table above. It's also, on its own, still a single flat rule applied bet after bet with no sense of where you are in a longer run.

A Masaniello cycle works differently again. Instead of one stake rule repeated forever, it works across a defined series of bets, a fixed number of selections with a target number of wins needed from that series. The stake for each individual bet in the cycle isn't fixed as a flat percentage either. It's recalculated as the cycle plays out, based on the bank remaining and how many of the target wins are still needed from how many bets are still left. Win early in the cycle and the remaining stakes adjust down, because fewer wins are now needed from the bets left. Lose early and the system has to work harder from what's left, which the stake sizing reflects directly rather than just carrying on blind.

The practical difference from plain percentage staking: a percentage plan reacts bet by bet with no memory of a target. A Masaniello cycle reacts to progress toward a defined goal across a defined stretch of bets, which is a different kind of risk control, not just a different formula.

Two Real Cycles, Bet by Bet

Numbers beat theory. Here's the same 13-bet cycle structure our own calculator ran, needing 7 wins at average odds of 3.0, worked from a £1,000 bank.

Example 1 — target hit

Seven wins landed by bet 10, so the cycle stopped there. Three bets in the original 13-bet plan never got used. Final bank £9,658.20, up £8,658.20 on the £1,000 start. Notice the stake shrinks after bet 6, not because of caution, but because six wins were already banked and the target needed only one more from the remaining bets. The system had already done most of the job it set out to do.

Example 2 — target missed

Same setup. Same £1,000 bank, same 13-bet plan, same 7-win target, same 3.0 average odds. Different, illustrative run of results.

Eight bets in, only 2 wins landed against 4 needed for the target still to be reachable, 5 wins required from 5 bets remaining, and one of those 5 was already lost. The cycle stops itself here rather than running the remaining bets pointlessly. Final bank £39.75, down £960.25 on the start, a 96% loss on the cycle.

Put the two side by side and the honest takeaway is this: the stake sizing is doing exactly the same job in both examples, adjusting to what's needed given what's left. What it can't do is make the results turn up. A string of early losses in Example 2 made 7 wins from 13 bets a genuinely hard ask, and the maths reflected that faithfully all the way down to zero reachability, rather than papering over it.

What Actually Pushes Risk of Ruin Up

The maths above assumed one thing that real betting rarely holds to: a fixed stake, kept fixed, forever. In practice, risk of ruin rarely blows up because someone sat down and decided to stake 20% a bet. It creeps there, one decision at a time.

Chasingis the biggest one. Lose three in a row, and the next stake quietly goes up, not because the analysis changed, but because the loss needs recovering faster. That single move can turn a controlled 2% plan into a 6% or 8% bet without the bettor ever consciously "deciding" to raise their staking percentage.

Stake creep during a winning rundoes the same damage from the other direction. Win a few, feel sharper than usual, and stakes drift upward while the underlying edge hasn't moved an inch. The bank is bigger, sure, but the stake grew faster than the bank did, which quietly raises risk of ruin exactly when it feels safest to ignore it.

Treating bonus funds as separatefrom the "real" bankroll is another common one. A £50 free bet doesn't sit outside your risk profile just because it didn't come from your own pocket. If it gets staked at a size that would be reckless with your own money, it's reckless, full stop.

No separate betting accountis a quieter version of the same problem. When betting money mixes with everyday spending, there's no fixed number to calculate a percentage against. The stake becomes "whatever's in the account today," and risk of ruin can't be meaningfully calculated against a number that isn't fixed.

Staking based on how a selection "feels" rather than what the calculation says is really the root cause behind every item on this list. Every one of these mistakes is a version of letting something other than the bank's actual size and the actual stake decide the stake.

How to Check Your Own Number

None of this needs to stay theoretical. Working out your own risk of ruin takes three numbers you already know.

First, your actual bankroll, the fixed amount set aside strictly for betting, not whatever happens to be sitting in the account. If that number isn't fixed, sort that first, everything downstream depends on it.

Second, your stake size, in £ per bet or as a percentage of the bankroll. If it changes bet to bet depending on confidence, use the largest stake you'd typically place. That's the one that actually determines your risk.

Third, an honest win rate. Not the number that feels right, the number your actual results support. Twenty or thirty settled bets is thin evidence for this. A hundred or more starts to mean something. Anyone without a real sample yet should use a conservative estimate rather than an optimistic one, since risk of ruin punishes an overestimated edge hard.

With those three, a risk of ruin calculator does the rest in seconds rather than by hand. If bets aren't flat, even-money wagers but a structured sequence with a win target, that's a different calculation entirely, the kind a Masaniello cycle is built to run, and the section above shows exactly what that output looks like bet by bet.

A Straight Note on Risk

Everything above is about managing risk that already exists. It isn't a way to make betting risk-free, and it isn't a substitute for treating betting as entertainment rather than income. No staking plan, including a Masaniello cycle, guarantees a win. What it controls is how much a losing run costs you, not whether one happens.

Bankroll management only works as a bankroll you can genuinely afford to lose. If betting is being used to try to solve a financial problem, or stakes are increasing to chase money already lost, that's worth stepping back from, not staking around. Free, confidential support is available through GambleAware and the National Gambling Helpline on 0808 8020 133.

FAQ