Flat Stakes vs Percentage Staking: Backing Favourites Over Time
I ran real sequences of favourites through flat and percentage staking. The plan that sounds safest did the most damage, because no staking method can rescue a negative return.
- 5median end bank, 5% of bank, from 100
- 96.9%median worst drawdown at 5% staking
- −3.6%flat-stake return on the bets themselves
Readers of my banker bets for today page often tell me percentage staking is the safe way to back short-priced favourites, because you can never go to zero. I tested that idea on real data. I took every favourite priced 1.50-1.99 across 22 European divisions from 2016/17-2025/26, 25,528 bets in total. Each league was played in date order from a start bank of 100 units.
The headline answer is uncomfortable. Flat 1-unit stakes left a median bank of 61 units, while staking 5% of bank left a median of 5, and every one of the 22 leagues finished down. The bets themselves returned −3.6% at a major bookmaker. That is the key: staking decides how fast a negative edge eats your bank, not whether it does.
01Flat and percentage stakes on identical favourites, compared
Every row in this table uses exactly the same bets. Only the stake changes. Flat 1 unit finished on a median 61 with a 50.3% worst drawdown, and 18 of 22 leagues ended below the starting 100. Tripling the flat stake to 3 units took the median bank to 0, with a 100.0% drawdown.
The percentage plans look gentler on paper but performed worse than the smallest flat stake. At 2% of bank the median ended on 40 and 20 of 22 leagues finished down. At 5% of bank the median was 5, and not a single league finished in profit.
| Staking plan | Median end bank | Median worst drawdown | Leagues finishing down | Median end bank (best price) | Down (best price) |
|---|---|---|---|---|---|
| Flat 1 unit | 61 | 50.3% | 18 of 22 | 91 | 12 of 22 |
| Flat 3 units | 0 | 100.0% | 18 of 22 | 73 | 12 of 22 |
| 2% of bank | 40 | 69.7% | 20 of 22 | 74 | 13 of 22 |
| 5% of bank | 5 | 96.9% | 22 of 22 | 29 | 15 of 22 |
The best-price columns matter just as much. Taking the highest price available lifted the flat-stake return from −3.6% to 0.0%, and the flat 1-unit median rose to 91. Better prices narrowed the gap but never reversed the order. Even at best prices, 5% of bank still left a median of only 29.
02How the median bank decays through each league's sequence
This chart follows the median bank as each league's run of bets plays out. All three lines drift down, which is what a losing return does over time. The speed of the fall differs a lot. Halfway through, flat 1 unit stood at 81, 2% of bank at 64 and 5% of bank at only 25.
The 5% line drops hardest early and never claws its way back. That is compounding working against you. After a losing run you are staking a smaller slice of a smaller bank, so the winners that follow cannot repair the hole.
View the data: median bankroll through the sequence
| Flat 1 unit | 2% of bank | 5% of bank | |
|---|---|---|---|
| 0% | 100 | 100 | 100 |
| 10% | 95.6 | 89.7 | 70.7 |
| 20% | 97 | 91.1 | 68.7 |
| 30% | 92.3 | 82.5 | 47.3 |
| 40% | 80.1 | 63.6 | 24.8 |
| 50% | 81.2 | 64.2 | 25.4 |
| 60% | 75.7 | 57.3 | 16 |
| 70% | 77.1 | 57.1 | 14 |
| 80% | 73.3 | 51.5 | 9.8 |
| 90% | 68 | 45.1 | 6.8 |
| 100% | 61.1 | 40.3 | 5 |
The usual comfort is that percentage staking cannot reach zero. In theory that holds, but the bank still fell below 1 unit, which is effectively bust, in 5 of 22 league sequences at 5% of bank. At 2% that happened in 0 of 22. Flat 3 units went bust in 15 of 22.
03Peak-to-trough falls show what each plan feels like
End banks tell you where you finish. Drawdown tells you what you have to sit through on the way. Flat 1 unit had a median worst fall of 50.3%, and 2% of bank had 69.7%. Then the bars jump to 96.9% at 5% of bank and 100.0% at flat 3 units.
A 96.9% drop means the account is technically alive but practically finished. Very few people keep staking calmly through that. I would rather judge a plan by its drawdown before I ever look at its best-case ending.
View the data: worst drawdown by staking plan
| Median peak-to-trough fall across 22 leagues, major-bookmaker odds | |
|---|---|
| Flat 1 unit | 50.3% |
| Flat 3 units | 100% |
| 2% of bank | 69.7% |
| 5% of bank | 96.9% |
Keep the spread in mind as well. With flat 1-unit stakes the best league, League One, ended on 125, and the worst, 2. Bundesliga, ended on 0. One caveat: these sequences settle same-day matches one after another, whereas real bettors stake several at once, which shifts percentage results slightly. And past sequences are no promise of future ones.
Value & Staking checklist
- Check the price first. Favourites at 1.50-1.99 returned −3.6% at a major bookmaker but 0.0% at the best available price. Without that price edge, no staking plan saves you.
- Keep stakes genuinely small. Flat 1 unit went bust in 2 of 22 leagues, while flat 3 units went bust in 15 of 22. Stake size drove ruin far more than the choice of method.
- Distrust percentage-staking comfort. At 5% of bank all 22 of 22 leagues finished down, with a median 96.9% drawdown. Never reaching zero is not the same as being protected.
Quick answers
Is percentage staking safer than flat staking on favourites?
Not in this study. Staking 2% of bank left a median of 40 units from 100, compared with 61 for flat 1 unit, and 5% of bank left just 5.
Do favourites priced 1.50-1.99 make money long term?
Across 25,528 bets the flat-stake return was −3.6% at a major bookmaker and 0.0% at the best available price. They broke even at best, and only with top prices.
How big a stake can a 100-unit bank handle on favourites?
Flat 1 unit had a median worst drawdown of 50.3%, while flat 3 units hit 100.0% and went bust in 15 of 22 leagues. Small flat stakes held up far better.
I'm Renata Salcedo, based in Los Cabos, and I write the best-bet selections at soccerbet.tips — the highest-confidence calls of the week, framed honestly around value and staking rather than empty guarantees.
View profile →
