The bankroll is the strategy
I went through a stretch in 2022 where I won 17 of 23 UFC bets and ended that stretch with less money than I started it with. The wins came on short favourites – small returns, decent hit rate. The losses came on three confident underdog plays that I’d sized too aggressively because the recent form felt good. Each underdog loss wiped out four or five of the favourite wins. The hit rate was excellent. The bankroll was decimated.
That sequence taught me something that no amount of reading about expected value had taught me: bankroll management isn’t a topic adjacent to betting strategy. It is the betting strategy. The selections matter, the odds matter, the read matters – but if the staking framework around them isn’t built to survive variance, then good selections lose money and bad selections lose more money. This piece is about the staking maths that actually works for UFC betting.
Why UFC is harder to bankroll-manage than most sports
Football betting variance is bounded by the fact that most football bets resolve within 90-120 minutes and the goal-scoring rate is roughly known. Tennis variance is bounded by the structure of the match – three or five sets, with mid-match recovery possible. UFC variance is different because individual fight outcomes have heavy tails in both directions.
UFC favourites win 68.12% of fights historically. The 2024 favourite win rate was 72% – meaningfully higher, suggesting either a tighter year for upsets or a structural shift in how the books are pricing favourites. Either way, the long-run hit rate on favourites is roughly two-thirds, and the average favourite price means winning two of three barely covers stake. The variance on favourite betting is real and the bankroll has to absorb runs of consecutive favourite losses without breaking.
Underdog betting has different variance again. Underdogs win 31.88% of fights, but the distribution is bimodal – some weight classes and matchup types produce upsets at notably higher rates while others produce them rarely. Bankroll-managing underdog plays requires accepting long runs of losses interspersed with occasional large wins. The biggest underdog winner in UFC history was Shana Dobson at +950 American odds against Mariya Agapova in 2020 – those wins are real but spaced widely, and the bankroll has to be sized to survive the gaps.
The fixed-unit model and why it works
The simplest staking model is fixed-unit betting. Pick a unit size – typically 1-2% of your starting bankroll – and bet exactly one unit on every selection, regardless of confidence level, regardless of price. The maths is elementary, the discipline is straightforward, and the bankroll variance is well-controlled.
On a £1,000 bankroll with a 2% unit size, every UFC bet is £20. Lose ten in a row, you’ve lost £200 – painful but not catastrophic. The bankroll is at £800 and the unit drops to £16 if you recalculate against the current bankroll, or stays at £20 if you fix the unit against the starting bankroll. Both approaches have defenders, but the fixed-against-starting approach is generally cleaner because it prevents the unit from spiralling down during a bad run.
The strength of fixed-unit betting is that it removes confidence-based sizing decisions. Confidence is a poor predictor of outcome on individual UFC fights – the matchup that feels obvious before the fight is often the one that resolves unexpectedly. Sizing equally across all selections forces the bettor’s edge to come from selection quality, not from “knowing which ones to bet bigger”. The 2-3% of bettors who consistently profit on UFC betting tend to use fixed-unit or near-fixed-unit staking, not confidence-weighted staking.
The proportional model and where it goes wrong
Proportional staking – bet a percentage of current bankroll on each selection, with the percentage scaled by some measure of edge – sounds mathematically elegant. The Kelly criterion is the canonical example, where stake size is proportional to edge divided by odds. In theory, Kelly maximises expected log-bankroll growth.
In practice, Kelly applied to UFC betting tends to over-stake, because the edge estimate is almost always wrong. Kelly assumes you know your edge – that you can accurately estimate that you have, say, a 4% expected value on a particular bet. Real UFC bettors estimate edges with substantial uncertainty. You might think you have a 4% edge; the true edge might be 1% or might be -1%. Kelly applied to overconfident edge estimates leads to over-staking, which leads to bankroll volatility that can wipe out an edge that’s genuinely there.
The fractional Kelly variant – bet 25-50% of what full Kelly would suggest – partially solves this. The reduction in stake size cushions against edge mis-estimation. The downside is that the same fractional adjustment that fixes the over-staking also reduces the expected growth rate, and the result is a system that’s structurally similar to fixed-unit betting but with more complicated maths and similar variance characteristics. For most bettors, fixed-unit is the simpler implementation of essentially the same risk profile.
The recovery rule that almost everyone gets wrong
The single most destructive habit in UFC bankroll management is increasing stake size after losses to recover the deficit. The logic feels intuitive – “I’m down £200, if I bet £50 on the next one and win I’m back” – and the structure is identical to gambling at any other game. Variance doesn’t reward recovery betting. It punishes it.
The mathematical reality is that your expected value on each bet is independent of your bankroll history. A £20 bet at +EV has the same expected value whether you’re up £500 or down £500 from your starting point. Doubling the stake to £40 on a “recovery bet” doubles both the expected value and the variance. If the bet is genuinely +EV, the long-run effect is positive; if it’s neutral or negative EV – which is more common when the bet is being selected to recover losses rather than for its own merits – the doubled stake doubles the rate of bankroll decay.
The rule that works is the inverse. After a loss, bet the same unit size. After a win, bet the same unit size. After ten losses in a row, bet the same unit size. The bankroll absorbs the variance, the selections find their long-run hit rate, and the discipline of consistent stake sizing is what allows the edge to compound.
The structural rules I apply to my own UFC bankroll
The first rule is that the bankroll is segregated from regular cash. The amount sitting in operator accounts and earmarked for UFC betting is money I can afford to lose entirely without affecting any aspect of my regular life. The bankroll is not topped up from regular income on a monthly basis. If it goes to zero, it stays at zero until I make a deliberate decision to recapitalise.
The second rule is that the unit size is 1.5% of the bankroll, reviewed quarterly. Every three months I look at where the bankroll stands and reset the unit size against the current balance. If the bankroll has grown, the unit grows proportionally. If it’s shrunk, the unit shrinks. The quarterly cadence is slow enough that short-term variance doesn’t push the unit around erratically.
The third rule is that no single fight gets more than 2 units of total exposure across all markets. If I’m betting the Moneyline at 1.5 units and want to add a method of victory bet on the same fighter, the method bet is capped at 0.5 units. The constraint prevents fight-specific variance from compounding through correlated bets on the same outcome.
The fourth rule is that I don’t bet on cards I haven’t watched at least one prior fight of for each fighter. This isn’t a bankroll rule strictly, but it functions as one – it prevents me from betting on fights I have no real read on, which would be variance with no edge component.
The drawdown numbers worth internalising
Even a sharp UFC bettor with a 55% hit rate at fair pricing will experience long drawdowns. The maths of variance means a 10-bet losing streak is not unusual; a 15-bet losing streak happens occasionally; a 20-bet losing streak is rare but not impossible. The bankroll has to survive those streaks without breaking.
With a 1.5% unit on a £1,000 bankroll, a 15-bet losing streak takes you to £775 – uncomfortable but workable, with the unit still at £15 (if fixed against starting) or recalculating to £11.60 (if fixed against current). A 20-bet streak takes you to £700, which is the point where most bettors emotionally start considering changes to the system. Either of those streaks happens to most UFC bettors at some point in their first year. The bettors who survive are those who knew to expect the drawdown and didn’t react to it.
The flat-out rule: if your drawdown depth is bigger than what your written staking rules would survive, your starting unit was too big. The fix is to halve the unit size and accept that the recovery will take longer. The fix is never to bet bigger to recover faster.
How operator stake limits intersect with bankroll planning
UK operators apply stake limits based on internal account assessments. A new account is typically capped low – £200-500 per UFC Moneyline bet – and the cap rises with betting history, deposit patterns, and the operator’s risk team assessment. Sharp bettors who consistently take +EV bets may find their limits decrease over time as the operator restricts them, in which case the bankroll has to be distributed across multiple operator accounts to maintain the planned unit size.
For most recreational bettors, operator limits are well above the 1-2% unit size on a typical recreational bankroll. A £1,000 bankroll with a £15-20 unit is well within any UK operator’s recreational caps. The constraint binds for larger bankrolls – a £20,000 bankroll with a £300 unit is bumping against the lower end of some operators’ caps on smaller markets, and the planning has to account for which markets accept the planned unit at which books.
The May 2025 UKGC affordability and financial vulnerability framework also intersects with bankroll planning. Operators conducting financial vulnerability checks more systematically since February 2025 will sometimes restrict stake limits or require source-of-funds documentation when bankroll-funded deposit patterns become significant. The framework is intended to protect vulnerable customers, but it affects sharp bettors with large bankrolls as well, and the operational friction is real. The next layer down from raw bankroll structure is how the unit size translates into actual bet sizing on specific markets – the staking unit conversion to UFC market-specific bets has its own logic.
