Bankroll Calculator: What a Measured Margin Does Over Many Bets
The register prices a single bet. This prices a season of them. Set a bookmaker, a stake and a number of bets, and the simulator runs the whole sequence a thousand times to show the range of places a balance can end up — and how little the staking pattern moves the average.
The simulator
Every run is seeded, so the same settings always produce the same figures and anyone can reproduce them.
—
middle 50% of runs (box) · middle 90% (line)
Why the staking pattern changes the picture but not the average
Run the simulator on flat staking, note the average loss, then switch to doubling after a loss and run it again. The distribution will look completely different: many more runs finishing slightly ahead, and a tail of runs that lost everything. The average loss will be the same figure, near enough, and it will match the margin multiplied by everything staked.
That is not a quirk of this implementation. Over any stopping rule, the expected loss equals the charge multiplied by the total amount staked. A staking pattern decides how much gets staked and in what order, so it moves the spread, the shape and the chance of ruin. It cannot move the expectation, because the charge is applied to every pound that passes through regardless of the order the pounds arrive in.
The two figures are printed side by side above for exactly this reason. Watching them agree across every pattern in the list is the most direct demonstration of the point the register is making one bet at a time.
The one input that genuinely changes the outcome is the margin itself, which is why it is selected from measured operators rather than typed in. Running the same settings against the keenest and dearest bookmakers in the study is the clearest illustration of what the price difference is worth over time.
What the simulation assumes, and where those assumptions break
Every run above treats bets as independent events settled at the margin you selected. That is the right model for the question being asked — what a known charge does to a balance over time — and it is the wrong model for almost everything else about a real betting account. The gap between the two is worth stating plainly, because a simulator that hides its assumptions is a worse tool than no simulator.
It assumes every bet carries the same margin. Nobody bets that way. The measured figures on the register are averages across 5 competitions, and the per-competition tables on each operator page show spans of several points inside one book. A run priced at an operator’s average describes an account spread evenly across its markets, not one that only ever backed the competition where that operator happens to be dearest.
It assumes a flat stake and an unlimited book. Real accounts meet maximum stakes, restricted markets and closed accounts, and none of those appear here. An operator that limits a winning customer changes the distribution of outcomes in a way no margin figure captures, and it is one of the reasons price alone is a partial description of an account.
It assumes the price is the only charge. Best odds guaranteed, price boosts and free bets all move the effective figure, usually downwards and unevenly — the modelled example puts a number on the first of those. None of them are simulated here, so treat the output as the cost of the margin by itself rather than as a forecast of a season.
What survives all three caveats is the shape of the result: the staking pattern moves the spread of outcomes a long way and moves the average almost not at all. That is arithmetic rather than modelling, and it holds however the assumptions above are relaxed.
Questions about the bankroll simulator
What does this bankroll calculator do?
It simulates a run of bets many times over and reports the distribution of where the balance ends up: the median, the spread, and how often the balance runs out entirely. Every run uses a seeded random generator, so the same settings always give the same result and anyone can reproduce it.
Where does the margin figure come from?
The bookmakers in our August 2026 study, so the simulation is driven by measured prices rather than a round number. Selecting an operator sets the per-bet charge to the margin actually measured on its match-result markets.
Does a staking system change the outcome?
It changes the shape of the distribution, not its centre. Doubling after a loss produces many small wins and rare very large losses; flat staking produces a tighter spread. The average loss stays at the margin multiplied by everything staked, for every pattern, which the simulator shows directly by printing both figures side by side.
Is this a staking strategy recommender?
No. The strategies are included precisely so the result can be compared across them and seen not to differ in expectation. Nothing here suggests one is better, because on this measure none of them is.
What is probability of ruin?
The share of simulated runs in which the balance reached zero before the run finished. It rises sharply with bet size relative to bankroll, and with staking patterns that increase stakes after losses.
Why does the average loss match the margin so closely?
Because that is a mathematical result rather than a coincidence: over any stopping rule, the expected loss equals the charge multiplied by the total amount staked. The simulation converges to it, and printing both numbers together is the clearest way to show it.
Can I use this to plan how much to bet?
It will show you what a given stake size does to the range of outcomes and to the chance of losing everything, which is useful information. It cannot tell you what to stake, and no simulation can make an expected loss into an expected gain.