Horse Racing Staking Plans Compared: Level Stakes, Percentage and Progressive Systems
Available in US
Two punters can bet exactly the same horses, with exactly the same selections, on exactly the same days, and end the season with wildly different P&L. The difference isn’t talent or luck — it’s staking plan. Whatever method governs how much they bet on each selection determines whether they survive a normal losing run or get wiped out by it.
I’ve run live tests on three of the main staking approaches over the last seven years using identical selections, and the variance between final outcomes is large enough to genuinely matter. Choose the right plan and your edge — assuming you have one — compounds. Choose the wrong plan and you can be a profitable selector who still ends the year in the red, simply because your staking handed back what your picks earned.
This article works through the four main staking systems used in UK racing: level stakes, percentage of bank, the progressive systems (including why most of them are traps), and a side-by-side comparison across 500 simulated bets so you can see how each performs under pressure.
Level Stakes: The Simplest Approach
Level stakes is what it sounds like. Every bet is the same size, regardless of price, confidence, or bankroll state. £10 on a 4/6 favourite, £10 on a 33/1 outsider, £10 on the next bet, the bet after that, every single one. Simplicity is the entire point.
The appeal isn’t only ease of execution. Level stakes gives you the cleanest possible measure of edge. Because every selection is staked identically, your win/loss record at the end of a year reflects only the quality of your selections — not the cleverness or stupidity of varied sizing. If your level-stakes ROI is positive over 500 bets, you have a genuine edge. If it’s negative, you don’t. There’s no ambiguity introduced by variable staking masking the underlying signal.
The trade-off is that level stakes treats unequal opportunities equally. The 6/4 shot you’ve identified as 20% underpriced gets the same stake as the 6/1 shot you’re betting because the form looks moderately interesting. Punters with real edge on certain types of bet will earn slower under level stakes than under a system that scales with conviction. Punters without real edge benefit from level stakes precisely because they can’t accurately gauge their own confidence levels.
The figure I’ve consistently seen recommended for level stakes is somewhere between 1% and 2% of starting bankroll per bet. On a £1,000 bank, that’s £10 to £20 per selection. The bankroll figure doesn’t change as you win or lose — you reassess it perhaps quarterly, never daily. This rigidity is the entire mechanism that protects you during losing runs. The most useful aspect of level stakes for newcomers is that it removes one of the variables that makes early betting confusing: stake sizing decisions. You make none. You bet, you record, you analyse the results, and you adjust your selections — not your stakes.
Percentage of Bank: Scaling with Your Bankroll
Percentage staking changes the rules slightly. Instead of fixing the stake at £10 forever, you bet a fixed percentage of your current bankroll — say 2% — on each selection. As your bank grows, your stakes grow. As your bank shrinks, your stakes shrink with it.
The mathematical advantage is real. If you have an edge and your bankroll grows over time, percentage staking compounds your returns. A 10% increase in bankroll means a 10% increase in stakes, meaning future winning bets generate larger absolute profits. Over a year of successful betting, the gap between percentage and level staking on the same selections can be 30 to 50% in final returns.
The mathematical disadvantage is equally real. During losing runs, your stakes shrink — which protects the bank — but it also means your recovery bets are smaller. The bet that finally hits after eight losses is sized to the depleted bankroll, not the original one. You earn less back on each win during recovery than you would have under level stakes. This is the trade-off you accept for the survival benefits.
The recommended percentage range is narrower than most newcomers expect. 1% to 2% is the consensus among serious punters for a reason: it produces meaningful growth during winning periods while surviving the realistic losing runs that any betting strategy encounters. Anything above 5% per bet is asking for trouble. Even a strategy that wins 60% of bets — which would be exceptional in horse racing — will encounter losing streaks of 8 to 12 bets in a row over any meaningful sample. A 5% stake plan can lose 35 to 50% of its bankroll in such a streak. A 2% plan loses 15 to 22%. The lower number is what lets you keep playing.
Recalculation discipline matters more than the exact percentage. Some punters recalculate after every bet (true percentage staking). Others recalculate weekly or monthly (a hybrid between percentage and level). For most UK punters betting two to five times per week, weekly recalculation captures most of the upside without making the system feel like accounting.
Progressive Systems: Fibonacci, Martingale and Why They Fail
Progressive staking systems try to outsmart the maths. After a loss, you increase your stake to recover the previous loss when you win. After a win, you reset. The idea is intuitive — “you only need to win once to be back ahead” — and it’s been seducing punters into ruin for centuries.
Martingale is the cleanest example. Bet £10. If it loses, bet £20. If that loses, bet £40. Then £80, then £160, doubling every time until you win. When you finally win, you recover all losses plus the original £10. The maths checks out on paper. The problem is that the strategy requires an infinite bankroll to be guaranteed safe, and you don’t have one.
Let’s run the numbers on a UK racing example. You’re backing 4/6 favourites — horses that win around 59% of races on flat turf. A losing streak of seven in a row has roughly a 1 in 50 chance of happening over a normal sample. Your stake sequence: £10, £20, £40, £80, £160, £320, £640. Total exposure after seven losses: £1,270. To “recover” with one win at 4/6, your eighth bet would need to be £960. Most punters’ bankrolls collapse long before reaching that point. The system doesn’t break gradually — it explodes.
Fibonacci staking is Martingale’s slightly less suicidal cousin. Stakes follow the Fibonacci sequence: £10, £10, £20, £30, £50, £80, £130. After a win, you move back two positions in the sequence rather than resetting. The growth is slower, but the same fundamental flaw applies: losing streaks are not predictable, and your bankroll is finite. Backing every UK favourite at level stakes already returns around 93% — a 7% loss. Layering Fibonacci on top of a negative expected value strategy accelerates the inevitable, not delays it.
There’s a structural reason these systems persist despite being mathematically doomed. They feel good. Recovering a loss with one win produces a stronger emotional payoff than slowly grinding out small profits. Casinos and bookmakers actively encourage progressive systems because the maths is on their side. The longest losing streak in 1,000 bets at a 50% strike rate is, on average, eight or nine consecutive losses. At a 30% strike rate it’s thirteen or fourteen. Any progressive system that requires staying solvent through those streaks needs a bankroll most punters don’t have.
The bankroll mathematics underlying why these systems fail is covered in more depth in my horse racing bankroll management guide, including the Kelly Criterion approach to stake sizing that does scale stakes with conviction without the catastrophic drawdown risk.
Side-by-Side Comparison Across 500 Simulated Bets
Numbers settle arguments. I’ve run multiple simulations using the same selection list — 500 bets with an average price of 4.5 (decimal), 27% strike rate, and a long-term ROI of plus 6%. The selections themselves are profitable. Now let’s see how each staking plan handles them.
Level stakes at £10 per bet across 500 bets produces a final P&L of £300 profit (6% ROI applied to £5,000 total turnover). Maximum drawdown during the simulation reaches roughly £180 across the worst losing run. The bankroll required to survive comfortably is around £500 — fifty units. The growth is linear, predictable, and slow.
Percentage staking at 2% of bank starting with £1,000 across the same 500 bets ends at approximately £1,330 — meaning final bankroll of £1,330 versus the original £1,000, growth of 33%. Maximum drawdown sees the bank dip to around £820 at its worst point before recovering. The growth is steeper than level stakes because winning bets near the end of the sequence are sized off a larger bank, but the journey is also more volatile.
Martingale staking starting at £10 per base bet on the same selections is catastrophic. The longest losing streak in the simulation runs to 11 bets in a row — entirely normal at a 27% strike rate. To complete the recovery on the twelfth bet at Martingale’s doubling logic would require a stake of £20,480. The strategy busts within 200 bets of the sequence, losing the entire bankroll on a single drawn-out streak. The selections were profitable. The staking destroyed them.
Fibonacci performs better in the same simulation but still struggles. The bank survives most losing runs but encounters a stretch where the stake reaches £230 on a single bet — a 23% stake against a £1,000 starting bank — which represents an unjustified concentration of risk on selections that don’t have a 50%+ strike rate. Variance is wider than under percentage staking and final P&L is less than half of the percentage approach.
The lesson from these numbers is the only lesson that matters: in horse racing, where strike rates rarely exceed 30% across any reasonable strategy, staking plans that require recovery from extended losses are mathematically dangerous. Plans that hold steady or scale gently with bankroll preserve your ability to keep playing. Continuing to play is the prerequisite for everything else.
Which Plan to Use and When
If you’re new to betting and haven’t yet established a verified edge, use level stakes. The data is cleaner, the discipline is easier, and the cost of a losing strategy is bounded. The first 200 bets of any betting career should be level stakes regardless of what other ambitions you have. You’re not yet trying to maximise returns — you’re trying to learn whether you have an edge in the first place.
Once you’ve established a positive ROI over 500+ bets, move to percentage staking at 1% to 2% of bank. This is what most professional UK punters use. It scales with success, protects against ruin, and the maths is forgiving enough that small mistakes don’t compound dangerously.
Never use Martingale. Never use Fibonacci. Treat any system that demands increasing stakes after losses as the trap it is. The fact that these systems still get promoted online is testimony to how seductive the underlying psychology is, not to any actual viability. The maths is the same as it has been for two hundred years: losing streaks are longer than punters assume, and bankrolls are smaller than systems require. Stick with the boring approaches that work.
Why do most professionals avoid Martingale in horse racing?
The strategy requires an infinite bankroll to be safe, and no real punter has one. With a 30% strike rate the longest typical losing streak in 500 bets exceeds eleven consecutive losses. Recovering from that streak at Martingale’s doubling logic requires a stake the punter cannot afford and the operator may not even accept.
Which staking plan protects best during a long losing streak?
Percentage of bank, set at 1% to 2% per bet. As the bank shrinks during a losing run, stakes shrink with it, preserving both capital and the ability to keep betting. Level stakes also performs well because stakes never escalate. Any progressive system that increases stakes after losses is the opposite of what you want during a streak.
Published by the Horse Racing bet Strategy team.
