This planner needs JavaScript. The method and a worked example are below the tool.
- Largest stake per bet
- —
- Chance a session runs out before its last bet
- —
- Budget per session
- —
- Expected loss per session
- —
- Likely session result (9 in 10 sessions)
- —
- Losing bets in a row the session budget absorbs
- —
How the planner works
A budget is first split evenly into sessions. Within one session the balance moves bet by bet: it drifts down by the house edge on average and swings around that drift by an amount that depends on how volatile each bet is. A session ends early if the balance can no longer cover the next stake, so the risk that matters is the chance of touching zero at any point, not only at the end.
The planner treats the balance as a random walk with a downward drift and uses the standard first-passage formula for it. With session budget B, stake s, house edge e, the standard deviation of one bet σ (in units of the stake) and n bets, the drift per bet is μ = e × s and the spread is σ × s. The chance of running out is
P = Φ((μn − B′) ÷ σs√n) + e^(2μB′ ÷ (σs)²) × Φ((−μn − B′) ÷ σs√n)
where Φ is the normal distribution and B′ is the budget less half a stake, because a player stops once the balance is below one full stake. The planner then searches for the largest stake at which P equals the risk you picked: one session in 20, in 10 or in 5. If a session is so short that even losing every bet cannot use up the budget, the stake is simply the budget divided by the number of bets, and the risk is shown as none.
A worked example
A $500 budget split into five sessions gives $100 per session. Planning 100 bets on red at European roulette (house edge 1/37, about 2.7%; standard deviation about 1.0) with a one-in-ten risk gives a stake of $5.22. At that stake the session can absorb 19 losing bets in a row, the expected loss is 100 × $5.22 ÷ 37 = $14.11, and nine sessions in ten should end somewhere between almost the whole $100 lost and about $72 ahead.
The figure was checked against a simulation of 30,000 sessions played bet by bet: at $5.22 about 9 sessions in 100 ran out before the hundredth spin. An earlier version of this planner only looked at where sessions ended and suggested $6.44 for the same inputs; simulated the same way, that stake runs out in roughly one session in five, twice the risk it promised.
How to read the results
- Largest stake per bet is a ceiling, not a target. Betting less only lowers the risk.
- Chance a session runs out repeats the risk at the rounded stake, so it can sit slightly under the level you chose.
- Expected loss per session is the house edge applied to everything you plan to wager: stake × bets × edge.
- Likely session result is the range nine sessions in ten fall into. It never goes below minus the session budget, because a session that runs out simply stops.
- Losing bets in a row shows how many consecutive losses the session budget covers at that stake, next to the number of bets you planned. Most sessions mix wins and losses, which is why the first number can be far smaller than the second.
Why the game changes the stake
Two games with the same house edge can need very different stakes. Red or black at roulette wins almost half the time, so results stay close to the average. A single-number bet wins once in 37 spins but pays 35 to 1, so the swings are about six times wider, and the same budget only supports a stake a fraction of the size. Slots sit in between or beyond, depending on the game's volatility; the planner uses a medium-volatility 96% RTP slot as its reference.
| Game | House edge | Volatility per bet |
|---|---|---|
| Blackjack, basic strategy | about 0.5% | low (σ ≈ 1.15) |
| Baccarat, banker bet | 1.06% | low (σ ≈ 0.93) |
| European roulette, red/black | 2.7% | low (σ ≈ 1.0) |
| European roulette, single number | 2.7% | high (σ ≈ 5.84) |
| Slots, 96% RTP | 4% | high (σ ≈ 6, varies by game) |
| Sports bet at even odds | about 2.5% | low (σ ≈ 1.0) |
With the same $100 session and 100 bets, the one-in-ten stake falls from $5.22 on red to about $1 on a single number, and on the reference slot the planner settles on $0.99, the level at which even 100 straight losses cannot empty the session.
What the planner cannot tell you
The random-walk formula fits many bets on low-volatility games closely. For single-number bets and slots it errs on the cautious side: a rare large win lifts the real chance of lasting the session, so the true risk is usually lower than shown. The planner assumes a flat stake throughout, and progressions such as doubling after a loss change the picture completely; the betting system simulator shows how. Blackjack doubles and splits put more than one stake at risk on a hand, which the reference figures only partly reflect.
No stake size changes the house edge. A plan decides how long the money lasts and how hard a bad session hits, which is why the budget should be money you can afford to lose. Setting the same figure as a deposit limit at the operator turns the plan into a rule the account enforces. If spending has gone past what you planned, the responsible gambling page lists free, confidential help.
About the author
Richard Smith
Founder & editor, GoodGamblingSites.com
Started the site to show where online gambling is legal. Every list here begins with the licence check.
- 12Countries & US states
- Licence firstRule for every list
About the site How sites are rated Reviewed 20 Oct 2026