Bankroll management controls how fast you can lose money; it can't change whether you're expected to lose it. A staking system — flat, percentage or Martingale — only reshapes your results. With a bookmaker's margin in every price, the average outcome of every system is still a loss.
That isn't a reason to ignore staking. A good plan keeps betting inside a budget you chose calmly, stops a bad evening from becoming a bad month and makes the cost of betting predictable. This guide shows how to set a bankroll, compares flat and percentage staking in a simulation, and explains why "progression" systems fail.
All numbers come from a fictional simulation run in Python. They describe a model, not real bettors.
Step one: set the bankroll
A bankroll is money set aside only for betting.
- It's money you could lose completely without affecting rent, bills, debt payments or savings.
- It's separate from your everyday crypto or bank balance — ideally in a separate account or wallet.
- You decide the amount in advance, not after a loss.
- You decide what happens when it's gone: you stop, and you don't top it up the same day.
- You set a time limit as well as a money limit.
If you can't tick all five, the right bankroll is zero.
The model we simulated
To compare staking plans fairly, we gave each one exactly the same kind of bet:
Starting bankroll: 1,000 units
Bets per run: 1,000
Price: 1.91 (implied 52.36%)
True win chance: 50.00% (fictional, coin-flip-like market)
Expected value: 0.50 × 1.91 − 1 = −0.045 per unit staked
Runs: 10,000 per plan (fixed random seed)1.91 on both sides of an even contest is a typical 2-way price with a margin of about 4.7%. The expected loss is 4.5% of every unit staked, whatever plan you use.
Flat vs percentage staking
- Flat staking: bet the same 10 units every time (1% of the starting bankroll).
- Percentage staking: bet 1% of the current bankroll every time, so stakes shrink after losses and grow after wins.
Results after 1,000 bets, across 10,000 runs:
| Result after 1,000 bets | Flat 10 units | 1% of current bankroll |
|---|---|---|
| Average bankroll | 553.2 | 635.7 |
| Median bankroll | 550.0 | 609.1 |
| Worst 5% of runs (5th percentile) | 9.8 | 370.6 |
| Best 5% of runs (95th percentile) | 1,046.6 | 982.1 |
| Runs that finished below 1,000 | 93.2% | 95.1% |
What the table shows:
- Both plans lost money on average. Flat staking lost about 447 units on average. That's close to the expected 1,000 bets × 10 units × 4.5% = 450.
- Percentage staking lost less in total — because it bet less. As the bankroll fell, so did the stakes, so less money was exposed to the margin.
- Percentage staking almost never wiped out. Its worst 5% of runs still kept about 371 units, while flat staking's worst 5% were down to about 10.
- Percentage staking finished below the start more often (95.1% vs 93.2%). Smaller stakes after losses make it harder to climb back.
So percentage staking is a better brake, not a better engine. It slows losses and protects the bankroll from ruin. It doesn't turn a negative expected value into a positive one.
Why Martingale fails
The Martingale system doubles the stake after every loss and returns to the base stake after a win. The idea is that one win recovers all previous losses plus a small profit.
We ran it on the same model: base stake 10, doubling on each loss, 1,000-unit bankroll, up to 500 bets.
Martingale, 10,000 runs
Average final bankroll: 734.2
Runs that hit a stake they could
not cover (bankroll too small): 89.0%The problem is exponential stakes. After six losses in a row the next stake is 640; after seven it's 1,280, more than the whole bankroll. Losing streaks that long feel rare, but over hundreds of bets they happen in most runs. When one does, the bettor loses a large share of the bankroll in a single sequence, which wipes out all the small wins before it.
Martingale also gets worse with real prices. At 1.91 a win after a losing streak doesn't quite recover the losses, because the profit on a win is only 0.91 times the stake.
Why no staking plan creates an edge
The underlying reason is simple arithmetic. Each bet has its own expected value. If every bet's expected value is negative, the total of any number of them is negative, no matter how the stakes are arranged.
Total expected result = Σ (stake on bet i × EV per unit)
If EV per unit < 0 for every bet, the total is < 0 for any stakes.Staking plans change three things: how big your swings are, how likely you are to go broke, and how long your budget lasts. They don't change the sign of the result. Only the price does — and to have a positive expected value you'd need to estimate probabilities better than a market that already includes the bookmaker's margin. Most people can't do that consistently, and short winning runs are not evidence that you can.
A practical staking plan
If you decide to bet, a plan like this keeps the cost visible:
- Set a monthly bankroll you can afford to lose.
- Use a small fixed stake, such as 1% of that bankroll, or 1% of what's left.
- Never increase stakes to win back losses.
- Set a stop for the session (for example, three losing bets or 30 minutes) and keep to it.
- Keep a simple log: date, market, price, stake, result.
- Review the log monthly. If the total is falling faster than you're comfortable with, stop.
If you're betting with crypto, decide your bankroll in a stable unit of value and see our crypto deposits guide for the costs that come before the first bet. And if betting stops feeling like entertainment, our responsible gambling page lists free, confidential support.
FAQ
What is a betting bankroll?
It's a fixed amount of money set aside only for betting, which you can afford to lose completely without affecting bills, savings or anything else.
Is flat staking or percentage staking better?
Neither creates profit. In our simulation both lost money on average. Percentage staking shrinks your stakes as the bankroll falls, so it loses more slowly and almost never empties the bankroll, but the average result is still a loss.
Does the Martingale system work for esports betting?
No. Doubling after every loss cannot turn bets with negative expected value into a positive total. In our simulation 89.0% of runs reached a point where the next doubled stake could not be covered.
How much should I stake per bet?
Only an amount you're comfortable losing. Many people use a small fixed slice of their budget, such as 1%, to make the budget last longer. That controls the pace of losses; it doesn't change the expected result.