No: a betting system that only changes how much you stake after wins and losses does not work. It cannot change the chance of any result or the price you are paid, so over any run of bets expected profit = edge x expected total staked, and a bet with a negative edge stays negative however it is sized.
With illustrative numbers, at $2.00 on a true 47.5% chance, every system below loses 5c per $1 staked on average. What a system changes is how much money goes through and how often a session ends ahead, which is why each one looks better over a short run than it is.
What each system does to your stakes
Every system counts stakes in units, such as $10 a unit, and changes the next stake after each result:
| System | After a loss | After a win | What it is trying to do |
|---|---|---|---|
| Martingale | Double the stake | Back to one unit | Recover every loss plus one unit with one win |
| Fibonacci | One step up the sequence 1, 1, 2, 3, 5, 8, 13, 21 | Two steps back | Recover losses more slowly than doubling |
| D'Alembert | Add one unit | Take one unit off, never below one | Profit when wins and losses even out |
| Labouchere | Add the lost stake to the end of a line of numbers | Cross off the first and last numbers | Finish the line for a set profit |
| Paroli | Back to one unit | Double, then back to one after three wins in a row | Ride winning streaks while risking little |
The martingale betting system has its own page, with the stake after each loss and the bankroll a long run needs.
Fibonacci betting
Each stake in the sequence is the sum of the two before it. After seven straight losses the eighth stake is 21 units and the run is 1 + 1 + 2 + 3 + 5 + 8 + 13 = 33 units down. A win moves two steps back, from 21 to 8, so one win does not clear a long run. Ten losses in a row cost 143 units, $1,430 at $10 a unit.
D'Alembert betting system
Stakes rise by one unit after each loss and fall by one after each win. The idea is that wins and losses balance out, but each result is independent of the last, so a run of losses is never owed a run of wins: that belief is the gambler's fallacy. Ten straight losses cost 1 + 2 + 3 + ... + 10 = 55 units.
Labouchere betting system
Write a line of numbers, such as 1-2-3 for a 6-unit target. Each stake is the first number plus the last. A win crosses both off; a loss adds the lost stake to the end of the line. When the line is empty, the session is 6 units ahead.
A win removes two numbers and a loss adds one, so a line can be cleared with fewer wins than losses, which is its appeal. The catch is the stakes: in the worked run below it put 104 units through to win 6.
Paroli betting system
The reverse of martingale: one unit after a loss, double after a win, and back to one after three wins in a row. A run can only lose its first unit, and three straight wins at $2.00 make 1 + 2 + 4 = 7 units. At a 47.5% chance that happens 0.475 to the power 3 = 10.7% of the time.
Why no staking pattern changes expected value
Each bet's expected profit = stake x edge, where edge = true chance x decimal odds - 1. A system sets each stake from results already in, before the next result is known, and the next result does not depend on the earlier ones. Add the bets up and a whole session obeys one rule:
expected profit = edge x expected total staked
At a true 47.5% chance on $2.00 bets, edge = 0.475 x 2.00 - 1 = -0.05. If a system puts $370 through a session on average, it loses 0.05 x $370 = $18.50 on average; one that puts $100 through loses $5. Raising stakes after five losses cannot beat this, because the sixth bet is still a 47.5% chance at $2.00. Four of the five systems raise stakes after losses, which is chasing losses written as a rule.
Only two things could break the rule: results that depend on earlier results, or stakes chosen after the result is known. Neither happens, which is why, over enough bets, the law of large numbers pulls every system towards the same -5% of turnover.
A system run over hypothetical results
Twelve hypothetical results at $2.00, run through each system: loss, loss, win, loss, loss, loss, win, win, loss, win, loss, win. That is 5 wins and 7 losses, and one unit is $10.
| System | Stakes, bet by bet (units) | Total staked | Profit |
|---|---|---|---|
| Level stakes | 1 on every bet | 12 units | -2 units (-$20) |
| Martingale | 1, 2, 4, 1, 2, 4, 8, 1, 1, 2, 1, 2 | 29 units | +5 units (+$50) |
| Fibonacci | 1, 1, 2, 1, 1, 2, 3, 1, 1, 1, 1, 1 | 16 units | 0 |
| D'Alembert | 1, 2, 3, 2, 3, 4, 5, 4, 3, 4, 3, 4 | 38 units | +2 units (+$20) |
| Labouchere (1-2-3) | 4, 5, 6, 6, 8, 10, 12, 11, 10, 14, 6, 12 | 104 units | +6 units (+$60) |
| Paroli | 1, 1, 1, 2, 1, 1, 1, 2, 4, 1, 2, 1 | 18 units | -6 units (-$60) |
The working for D'Alembert: its five wins came on stakes of 3, 5, 4, 4 and 4 units, winning 20; its seven losses were 1, 2, 2, 3, 4, 3 and 3 units, losing 18; profit +2 units.
Four of the six finish level or ahead on a sequence where level stakes lose. That comes from the order of the results, not from the systems: shuffle the same 5 wins and 7 losses and the ranking changes. Labouchere's +6 needed $1,040 staked in 12 bets, and Paroli lost most because its doubled stakes landed on losers.
Every system here also assumes even money, where a win's profit equals its stake. At an illustrative $1.90 a win's profit is 0.90 of the stake, so each system's target falls short before the margin is counted:
| Run | Stakes (units) | Result at $2.00 | Result at $1.90 |
|---|---|---|---|
| Labouchere 1-2-3: loss, win, win | 4, 5, 5 | +6 units | -4 + 4.5 + 4.5 = +5 units |
| Fibonacci: loss, loss, win | 1, 1, 2 | 0 | -1 - 1 + 1.8 = -0.2 units |
In a three-way soccer market a bet on either team also loses on a draw, so a system's stakes move as if that team had lost.
What every possible session shows
One sequence proves nothing, so count them all. Ten bets can land 1,024 ways. Weighting each way by its chance at a true 47.5% gives each system's exact averages, at $10 a unit, with Labouchere starting a new 1-2-3 line whenever one clears:
| System | Average staked | Average result | Result per $1 staked | Sessions ahead | Worst session |
|---|---|---|---|---|---|
| Level stakes | $100.00 | -$5.00 | -5% | 31.7% | -$100 |
| Martingale | $370.68 | -$18.53 | -5% | 81.9% | -$10,230 |
| Fibonacci | $167.86 | -$8.39 | -5% | 55.3% | -$1,430 |
| D'Alembert | $229.73 | -$11.49 | -5% | 60.4% | -$550 |
| Labouchere (1-2-3) | $573.71 | -$28.69 | -5% | 58.7% | -$950 |
| Paroli | $160.67 | -$8.03 | -5% | 44.8% | -$100 |
Every system's result per dollar staked is -5%, the edge on each bet: for martingale, -$18.53 / $370.68 = -0.050. Martingale finishes ahead in about four sessions in five and still has the second-worst average, because its losing sessions are huge. A system decides how often you finish ahead and how bad the worst session gets. It never decides the average.
What can change outcomes: price and selection
Only two things move expected value: the price you take and the true chance of what you back. With illustrative numbers:
| Bet | True chance | Price | Expected value per $1: chance x price - 1 |
|---|---|---|---|
| The bet above | 47.5% | $2.00 | -5.0% |
| The same selection at a better price | 47.5% | $2.15 | +2.1% |
| A selection with a higher true chance, at $2.00 | 51% | $2.00 | +2.0% |
A better price for the same bet, found by comparing bookmakers' prices, changes the value of every dollar staked under any staking plan. A better estimate of the chance is the harder route, and it only counts if your estimate is closer to the truth than the market's. Value betting covers both, and the expected value calculator runs the sum.
Staking still has a job in a betting strategy, just a different one: sizing bets so a bankroll survives losing runs. Staking plans compares the approaches that do that without claiming to create an edge.
Risk: Betting involves risk. A bet with positive expected value can still lose, and often does, and a true chance is always an estimate. There is no guarantee of profit. See responsible gambling for limits and support.
Red flags in systems sold online
Systems are sold as courses, software, spreadsheets and tipping services. Treat any of these as a warning sign:
- A promise of profit, a "win rate" with no prices beside it, or talk of investing and income.
- A recovery or progression staking plan as the main selling point.
- Results shown as screenshots or highlights instead of every bet, with its date and the price taken.
- A record that starts just after a winning run, or rests on a few dozen bets.
- Pressure to buy now, a countdown, or a second fee for "the full method".
- A seller who earns from selling the system rather than using it, or who goes quiet once you have paid.
Scamwatch, run by the ACCC, warns that sports "investment" schemes, also sold as sports arbitrage, betting, tipping or trading, are just another form of gambling and often outright scams. They usually involve buying prediction software or joining a betting syndicate, sometimes with a tipping subscription. Problems reported to the ACCC include software that was never delivered or did not work, money that vanished from betting accounts with no bets placed, and sellers who could not be contacted (checked October 2026).
Backtesting betting strategies shows how to check a seller's rule on past prices, bet by bet, instead of trusting their record. Fake tipster red flags covers the tipping side, and sports investment scheme scams the schemes.