Flat Staking vs Unit Sizing for NBA Bankrolls - UK Guide

The single best decision I made as a UK NBA bettor was deciding, in writing, what one “unit” meant before I placed my first bet of any given season. The reason is straightforward and slightly embarrassing – without that decision, I’d bet £50 on a game I felt strongly about, £20 on a game I was lukewarm on, and £100 on a Friday-night flier because the multiplier looked exciting. Whatever skill I had at picking games was being drowned by an undisciplined stake-sizing process. Andrew Rhodes of the UK Gambling Commission framed the wider importance of this kind of thinking when he noted: “The Gambling Survey for Great Britain is a key building block of the evidence base which helps government, industry and other partners understand both gambling behaviour and potential consequences from gambling.” Stake-sizing discipline is the difference between behaviour you can defend over a season and behaviour you can’t.

Two systems dominate the UK NBA bettor’s toolkit: flat staking and percentage-of-bankroll unit sizing. Both have advocates and both have legitimate use cases. The choice between them depends less on which is mathematically optimal – the maths is well-established – and more on your own discipline, conviction, and tolerance for variance.

What “a unit” actually means

A unit is your standardised betting denomination. It’s the amount you stake on a normal bet – not a high-conviction play, not a recreational flier, but the default amount for a typical NBA bet that fits your selection criteria. If your unit is £20, then a normal bet stakes £20. The whole stake-sizing system then revolves around how you size bets relative to that unit.

Setting the unit size is the first decision and the one most punters get wrong. The convention I’d recommend is 1% to 2% of your starting bankroll. If you’ve committed £1,000 to NBA betting for the season, a unit is £10 to £20. That feels small to people who came from football betting on bigger stakes, but it’s deliberately conservative. You’ll place a lot of NBA bets across a season, variance is real, and the 1% to 2% range absorbs losing streaks without forcing you to chase.

The UK adult problem-gambling rate sits at 2.7% (PGSI 8+) according to the most recent GSGB cycle, and that population skews toward higher-stake, lower-discipline behaviour. The 18-24 demographic is particularly exposed – 21.9% have PGSI scores between 1 and 27 – and unit sizing at the right level is one of the practical defences against the pattern of inflating stakes after losses. Setting a unit at 1% of bankroll and refusing to move it is exactly the kind of pre-commitment that protects against the worst behaviours.

The unit should not change based on recent results. If you’ve won six bets in a row, your unit stays the same. If you’ve lost six in a row, your unit stays the same. The only time the unit changes is when you formally rebalance – typically at the start of a new season or after a substantial bankroll change.

Flat staking mechanics

Flat staking is the simplest possible system: bet the same amount on every bet, regardless of conviction or price. One unit per bet, every bet.

The strength of flat staking is that it removes one decision from the table. You’re not asking yourself whether this is a 2-unit play or a 0.5-unit play; you’re staking one unit and moving on. That reduces the chance of letting recent results bias your sizing, which is one of the biggest leaks in unsystematic betting.

The mathematical case for flat staking is that, in the absence of reliable probability estimates, all your bets carry similar expected value. If you can’t quantify your edge per bet, you can’t size proportionally to edge. Flat staking is the honest acknowledgement that you don’t know which bets are higher-edge than others – and that’s an acknowledgement most NBA punters should be making.

The weakness is that you’re foregoing the value of varying stake by conviction when you do have reliable conviction. A genuine high-conviction play at the same stake as a marginal one wastes available bankroll on the marginal bet and underweights the high-conviction one. If your model is well-calibrated, you’re leaving money on the table.

The honest answer is that most UK NBA punters don’t have models well-calibrated enough to size proportionally. Flat staking is the safe default. If you want to vary stake sizing, the discipline required to do it well is much greater than the discipline required to flat-stake, and most punters don’t put in the work.

Percentage of bankroll versus fixed unit

Within flat-stake systems there’s a meaningful choice between fixed-unit staking and percentage-of-bankroll staking.

Fixed-unit staking sets the unit at the start of a period – typically a month, sometimes a season – and bets that amount until the period ends, regardless of bankroll changes. If you started January with £1,000 and a £20 unit, you bet £20 per bet through January even if your bankroll is at £1,200 or £800 mid-month.

Percentage staking recalculates the unit after every bet. If your unit is 2% of bankroll and you win, your bankroll grows and so does the next unit. If you lose, your bankroll shrinks and so does the next unit. The arithmetic compounds gains and dampens losses – exponentially good times and gentler bad times.

Both have their place. Percentage staking is mathematically superior in the long run because it scales with your edge across changing bankroll. Fixed-unit staking is psychologically easier because the stake doesn’t change after every bet, which removes the small temptation to bet bigger after winning streaks. UK punters who are honest with themselves about their discipline tend to favour fixed-unit for the same reason most people favour fixed mortgage rates – predictability matters more than optimisation when the alternative invites bad behaviour.

UK demographics for sports betting reinforce the discipline case. About 10% of UK adults placed sports or racing bets online or via an app in the April-July 2025 cycle, and within that population the distribution of stake-sizing approaches is wide. The bettors who report consistent results almost universally use a fixed system. The bettors who report wild swings almost universally don’t.

Which system suits which bettor

The choice between flat staking, fixed unit, and percentage staking depends on three variables: your edge confidence, your discipline, and your variance tolerance.

If your edge is uncertain – which is true for the vast majority of UK NBA punters – flat-stake fixed-unit is the right default. You’re not pretending to know which bets are higher-edge than others, and you’re not letting recent results push you around. The system is boring and that’s the point.

If your discipline is high and your edge is genuinely well-modelled – you’ve kept records for 500-plus bets, calculated CLV, and identified specific market niches where you consistently beat the close – percentage-of-bankroll staking starts to make sense. You’re getting compounding benefits from your edge, and your discipline protects against the recency-bias trap.

If your variance tolerance is low – meaning you’ll panic-withdraw after a 25% drawdown – go conservative on your unit size regardless of system. A 1% unit on flat staking means a 50-bet losing streak only takes 50% off your bankroll, which is a lot but survivable. A 5% unit on the same losing streak takes you to zero. The unit size is the lever that controls variance more than the system you use.

One layer I’d add for every UK punter regardless of system: deposit limits and time-outs as a structural backstop. The maths is one defence; the operator-side tools are another. The deposit limits and self-exclusion tools complement bankroll rules piece works through how those operator tools layer onto your stake-sizing system.