Bankroll and risk
A bankroll is the money set aside for trading that you could lose in full without it changing anything else in your life. This page covers how to count it honestly, the three ways traders decide what goes behind a single position, and why none of them can turn a losing approach into a winning one.
What a bankroll actually is
A bankroll is a ring-fenced sum: separate from rent, bills, savings and anything else the money is already committed to, held on your exchange accounts or set aside for them, and an amount you can afford to lose entirely. A figure that fails that description is money doing two jobs at once, and the second job will make the trading decisions.
Count it honestly. The bankroll is what is genuinely available now: the funds on your exchange accounts, less anything already tied up as liability on open positions. A deposit you haven't made isn't part of it, and neither is money you'd only add after a bad run.
Sizing a position
Every staking plan answers one question — how much of the bankroll goes behind this trade? — and they differ mainly in what they do as the bankroll moves.
Flat staking
The same amount every time. On a £500 bankroll, £10 a trade is 2% of the starting figure and it stays £10 whether the bankroll grows to £650 or falls to £400 — nothing to recalculate in a hurry, which is why it is the sensible starting point. The trade-off runs both ways: it does not compound as the bankroll grows, and it does not shrink during a losing run either, so £10 out of £400 is a bigger share of what is left than £10 out of £500 was.
Percentage of bankroll
A fixed fraction of whatever the bankroll is right now. At 2% of £500 the stake is £10. Lose a run of them and the bankroll falls to £450, so the next stake is £9; grow it to £650 and the stake is £13. Sizing falls automatically when things go badly and rises as the bankroll does, without you deciding anything in the moment. The catch is that recovery is slower than the decline, because a smaller stake has to win back ground a larger one lost — and exchange minimum stakes eventually put a floor under the fraction.
Kelly, and why people use a fraction of it
The Kelly criterion sizes a bet from the edge you believe you have. For decimal odds o and an assessed win probability p, the fraction of the bankroll is f = (p × (o − 1) − (1 − p)) ÷ (o − 1). At odds of 3.0 with p assessed at 0.40: o − 1 = 2, the numerator is (0.40 × 2) − 0.60 = 0.20, and f = 0.20 ÷ 2 = 0.10. Ten per cent of a £500 bankroll is £50 — five times the 2% stake above.
The catch is that Kelly assumes your probability is correct — and you supplied it. Break-even at 3.0 is exactly one in three, so watch for your estimate sliding towards it.
| Assessed probability | Kelly fraction | Stake on £500 |
|---|---|---|
| 0.400 | 10.00% | £50.00 |
| 0.375 | 6.25% | £31.25 |
| 0.350 | 2.50% | £12.50 |
| 0.333 | 0% | nothing |
Calling 40% something that is really 35% asks for four times the stake the true figure justifies, and at 33.3% the correct stake is nothing at all while the 40% you entered still says £50. Full Kelly also produces swings most people cannot sit through. Between them, that is why halves and quarters get used instead: half Kelly here is £25, a quarter £12.50.
On a lay bet, liability is the number that matters
Back a selection and the stake is what you risk. Lay one and it is not: a lay of stake S at odds O risks S × (O − 1), the amount you pay out if the selection wins.
| Lay stake | Odds | Liability |
|---|---|---|
| £20 | 2.50 | £30 |
| £20 | 6.00 | £100 |
| £20 | 11.0 | £200 |
| £20 | 21.0 | £400 |
Laying at 6.0 for £20 risks £100 — the same £20 that would have risked exactly £20 backed. Beginners size the stake and meet the liability afterwards. Apply the bankroll rule to the liability instead: 2% of £500 permits £10 at risk, and at odds of 6.0 that is a lay stake of £2, because £2 × 5 = £10.
Exposure is counted across markets, not within one
Several open positions on related outcomes are really one position. Backing a team in the match odds, backing the same team on a correct score and laying the draw is one event deciding all three: they win together and lose together, so splitting a stake across them spreads nothing. Two exchanges are not two bankrolls either. Total exposure — everything you lose if the event goes the wrong way, across every market and exchange — is what the rule applies to.
No staking plan creates an edge, and none prevents losses. Sizing controls how fast a bankroll moves, not which direction — applied to trades that lose money on average, it changes only how long the money takes to go. Nothing here is a prediction of what you would make, and every approach in this section can lose money. Never stake more than you can afford to lose.
Check your real exposure in Positions & P&L before sizing the next trade, not after: every open position across your connected exchanges is listed there, marked to live prices, on Basic as well as Pro. Then set your unit once in Preferences → General → Trading Defaults, which holds the default one-click stake every new market window starts with. BetTrading.net does not enforce a staking plan or a bankroll limit for you — keeping to it is yours.
Chasing losses is the failure this prevents
The thing a staking plan exists to stop is the raised stake after a loss — the trade sized to win back the last one rather than sized to the bankroll. A size you set in advance was chosen calmly. A size you set after three losses was chosen by the three losses. Review the record rather than the feeling: Trade history keeps every bet you have placed — 30 days on Basic, an unlimited range with CSV and PDF export on Pro — and a month of it shows what your sizing actually was.
Stepping away is a legitimate decision, not a failure of nerve. Closing the application for the evening costs nothing. And if gambling stops being fun, free and confidential support is available at BeGambleAware.org.
Decide the number before the session: write your unit size down before you open a market, in money rather than percentages, so no arithmetic happens while a price is moving. Recalculate it monthly, never after a result.