To beat the bookies you have to bet at prices better than the true chance of each result, often enough to cover the margin built into every market. Betting at random pays that margin on average: on a market whose prices add up to 105%, it gives away close to 5 cents of every dollar staked.
A short list of approaches can narrow or clear it: value betting against fair prices, taking the best price on offer, using promotions already on your account, and arbitrage and middles. Each has limits, and none of them makes the result of a bet certain.
Why punters lose to the margin
A bookmaker's prices imply chances that add up to more than 100%, and the excess is its margin. With illustrative prices for a two-team game:
| Measure | Team A | Team B | Total |
|---|---|---|---|
| Bookmaker price | $1.65 | $2.25 | |
| Implied chance (1 / price) | 60.61% | 44.44% | 105.05% |
| Fair chance (margin taken out) | 57.69% | 42.31% | 100% |
| Fair price (1 / fair chance) | $1.73 | $2.36 |
The fair chances come from the simplest way of removing a margin, scaling each implied chance down by the same factor: (1 / 1.65) / 1.0505 = 0.5769, or 57.69%. A $1 bet on either side then returns 1 / 1.0505 = $0.952 on average, an expected loss of about 4.8 cents. Over $10,000 staked at random on markets like this, that is an expected loss of about $481.
Multis compound it. Two independent legs from markets like this return 0.952 x 0.952 = $0.906 per dollar on average, and three return about $0.863: expected losses of about 9.4% and 13.7%. Bookmaker margin shows how to measure the margin on any market, and can you make money betting looks at what it means for punters in general.
Five approaches that narrow or beat the margin
| Approach | Where the edge comes from | Main limit |
|---|---|---|
| Value betting | A price above your estimate of the true chance | Variance, and a fair price that is only an estimate |
| Taking the best price | The highest price across several bookmakers | It beats the margin only when the best price is above fair; otherwise it just shrinks it |
| Promotions on your account | Bonus bets, refunds and boosts a bookmaker has issued | Capped value and each bookmaker's terms |
| Arbitrage | Two bookmakers' prices that add up to under 100% | Thin margins, and legs refused or voided |
| Middles | Two lines with a gap that both bets can win in | Loses on most results by design |
Account restrictions can cap any of the five.
Value betting against fair prices
A value bet is a price above the fair price, which can come from a market with its margin removed, from the exchange or from your own ratings; how to calculate fair odds compares the three. Edge = offered price / fair price - 1, so $5.50 against a fair $5.20 (illustrative) is 5.50 / 5.20 - 1 = 5.8%. If that fair price is right, 5.8 cents per dollar staked is also the bet's expected value, and the bet still loses 1 - 1 / 5.20 = 80.8% of the time.
On the rule in sample size in betting, a 5.8% edge at $5.50 takes about 4 x 4.5 / (0.0577 x 0.0577) = 5,400 bets just to reach two standard errors on average, and about twice that to clear them 4 times in 5. A fair price that is wrong turns value into a slow loss, and value betting covers how to test one.
Taking the best available price
Comparing prices before every bet attacks the margin directly. Suppose, with illustrative prices, a second bookmaker offers Team A at $1.72 and a third offers Team B at $2.32. The best prices add up to 1 / 1.72 + 1 / 2.32 = 101.24%, against 105.05% at one bookmaker. Measured against the same fair chances, the expected loss on Team A falls from 4.8% to 0.8% (0.5769 x 1.72 = 0.992), and on Team B to 1.8% (0.4231 x 2.32 = 0.982).
That is better, and still below zero: comparing prices shrinks the margin you pay, but beating it also takes a price above fair. Line shopping works through the long-run maths. Prices move, so confirm the price in your bookmaker account before you bet.
Promotions already on your account
Bonus bets, money-back refunds and bet boosts can carry positive expected value, because the bookmaker adds value to a bet that would otherwise pay the margin. Australian bonus bets normally return only the winnings. An illustrative $50 bonus bet at a fair price of $6.00 is a 1 in 6 chance of 50 x (6.00 - 1) = $250, worth $250 / 6 = $41.67 on average. How to value a betting promotion gives the method for each kind.
Matched betting can turn part of that value into a steadier result by laying the same selection on a betting exchange, if the lay is matched at the price you planned. The matched betting guide covers the method and its risks, and the bonus bet converter works out the lay stake and the cash you keep.
Under the National Consumer Protection Framework, a bookmaker cannot offer a reward or other benefit to get you to open an account, and any winnings from a bonus bet have to be withdrawable with no turnover requirement attached. Each bookmaker's terms decide which bets qualify, the value is capped by the offer, and a bookmaker can withdraw promotions from an account.
Arbitrage
An arbitrage backs every outcome of an event at different bookmakers, when the best prices add up to under 100%. With illustrative prices of $2.08 on one player at Bookmaker A and $2.00 on the other at Bookmaker B, the book is 1 / 2.08 + 1 / 2.00 = 98.08%. Split $500 as $245.10 and $254.90, and whichever player wins returns about $509.80: $9.80, or about 2% of the total staked.
That $9.80 needs both bets to stand, and a player's retirement can be settled differently at each bookmaker. The arbitrage betting guide covers those risks, and the arbitrage calculator works out the stake split and profit from your own prices.
Middles
A middle loses on most results by design. With illustrative prices, $50 on Team A -2.5 at $1.87 at one bookmaker and $50 on Team B +5.5 at $1.87 at another cost $100. Any result outside the gap wins one bet only and returns $93.50, a loss of $6.50. If Team A wins by 3, 4 or 5, both bets win and return $187, a profit of $87.
Break-even needs the gap to land 6.50 / (87 + 6.50) = about 7% of the time. The middle percentage, -6.5% here, is the loss per $100 staked on a miss: it is not a probability, and a middle is not an arbitrage. Read middle betting before you try one.
Risk: Betting involves risk. A +EV bet can still lose, and often does; an arbitrage or a middle only works if every bet is accepted at the price shown; and a bonus bet that loses pays nothing. There is no guarantee of profit. See responsible gambling for limits and support.
The limits: variance, restrictions, time and capital
| Limit | What it means in practice | Hits hardest |
|---|---|---|
| Variance | A real edge can sit behind a losing record for thousands of bets | Value betting, middles |
| Restrictions | Smaller maximum stakes, promotions withdrawn, accounts closed | Any approach that keeps winning |
| Time | Prices need checking across bookmakers close to the start, and gaps close as prices move | Arbitrage, middles, taking the best price |
| Capital | Money has to sit in several accounts, and an arbitrage ties up both stakes until it settles | Arbitrage, taking the best price |
The arbitrage example put $500 to work for $9.80. If one leg were refused after the other was placed, the open bet would carry the full risk of a single bet at its price, the problem called a one-legged arb.
What happens to an account that wins
A bookmaker earns its margin from the average customer, so an account that keeps beating the market's final price, bets mainly when a promotion adds value, or backs one side of an arbitrage costs it money on average. Under their terms, bookmakers can generally cut maximum stakes, withdraw promotions, void bets in the cases the terms set out or close the account, and many terms restrict automated betting, third-party access and multiple accounts. That risk is yours, and it lands differently on each approach:
| Approach | What a restriction takes away |
|---|---|
| Value betting | Stake size, so the same edge earns less per bet |
| Promotions | The offers, which can be withdrawn from the account |
| Arbitrage and middles | The second leg, if its stake is cut or refused after the first bet is placed |
| Taking the best price | Accounts to compare, so the best price open to you falls |
Why bookmakers restrict accounts covers the triggers, where to complain and the rights you keep. Among them are the minimum bet limits set by racing bodies in some states, such as Racing Victoria and Racing Queensland: the amount an off-course operator must accept an eligible fixed-odds racing bet to lose (checked October 2026). A restriction is not something to get around through someone else's account or identity, or through an offshore site.
Section 15AA of the Interactive Gambling Act makes it an offence to provide online betting to customers in Australia without a licence from an Australian state or territory. ACMA lists what can go wrong for punters: an illegal operator may not pay out winnings, and Australian regulators cannot help its customers.
What does not beat the bookies
| What people try | Why it does not beat the margin |
|---|---|
| Staking systems such as martingale | They change how much you bet, never the price, so a negative expected value stays negative; betting systems that don't work tests the common ones |
| Raising stakes to win back losses | More money goes through the same margin; chasing losses explains why it happens and where to get help |
| Bought systems, syndicates and tipping schemes | The ACCC's Scamwatch says sports investment schemes are not investment opportunities but just another form of gambling, and often outright scams (checked October 2026); see sports investment scheme scams |
How to tell whether you are beating the bookies
Four checks show whether any approach, or any betting strategy, is working:
- Record every bet, with the price you took and the closing price; how to track betting results lists the other fields.
- Work out your yield, profit divided by total staked, with cash bets and bonus bets kept apart, as ROI vs yield explains.
- Work out your closing line value on each bet, your price / the closing price - 1, with the margin taken out of the close. It shows an edge far sooner than profit does.
- Judge both by sample size: each needs to sit about two standard errors above zero, at your prices, before it means much.
Set a deposit limit with each bookmaker before you start testing: Australian online bookmakers must offer one, a decrease applies straight away, and an increase only after 7 days.
Where B337's screeners fit
B337's sports screeners search for candidates for three of these approaches: value, arbitrage and middles. The +EV screener lists bookmaker prices that beat a fair price, ranked by edge. It builds that fair price by stripping the margin out of prices across the market, so it is an estimate, not a known probability.
The arbitrage screener lists opposing prices that, on paper, return more than the total staked whichever side wins, and the middles screener looks for two lines with a gap between them.
The rows come with Terminal Pro or Full Automation, or with a Screeners plan of their own. On a free account you see how many candidates are live, but the rows stay locked, and positive EV betting shows how the +EV screener works. The screeners find candidates; each bet is yours to decide and to place, in your own bookmaker accounts, and B337 makes no claim that it avoids limits.
Risk: Betting involves risk. Prices move between the screen and your bet, a screener's edge is an estimate, and bookmakers can restrict accounts and void bets. See responsible gambling for limits and support.
For free and confidential support call 1800 858 858 or visit gamblinghelponline.org.au.