A 5% vig per bet becomes a 92% chance of losing over 500 bets. The arithmetic of why almost everyone loses — and what it actually takes to finish in the 2%.
Our AFL model is right 71% of the time and loses money. Our NRL try scorer model is right 26% of the time and prints profit.
The betting market knows more than most models give it credit for. We trained on 233K races to find out exactly where the edge lives — and where it doesn't.
Bookmakers do not need you to be bad at picking winners. They only need you to keep betting. The margin — the vig, the overround, the juice — does the rest. A 5% margin per bet does not feel like much. Over 500 bets, it becomes a 92% probability that you are losing. Over 1,000 bets, 99%. The math does not care about your sport knowledge, your system, or your winning streaks. It compounds silently, every bet, until almost everyone is underwater.
This piece walks through the arithmetic. No opinion. No strategy advice. Just the cold numbers that explain why approximately 98% of punters lose over the long term — and what the numbers say about what it takes to be in the 2% that do not.
Every bet you place at a corporate bookmaker includes a margin. On a standard two-outcome market priced at $1.91 each side, the book percentage is:
1/1.91 + 1/1.91 = 1.047 = 104.7%
A fair market (no margin) sums to 100%. The 4.7% above 100% is the vig. It means that if the true probability of each outcome is 50%, the fair price is $2.00. The bookmaker is paying you $1.91 — a 4.5% discount from fair value. That discount is the bookmaker's edge. Every bet you place at $1.91 on a true 50% outcome has an expected value of:
EV = 0.50 × $0.91 + 0.50 × (-$1.00) = -$0.045 per dollar staked
You lose 4.5 cents per dollar, per bet, on average. It does not matter whether that individual bet wins or loses. Over a large number of bets, the expected value resolves. The 4.5 cents per dollar accumulates into real losses.
Different bet types carry different margins. A head-to-head market might carry 4-5% vig. A three-outcome market (win/draw/win) might carry 6-8%. A multi-leg exotic might carry 15-30% vig embedded across the legs. A same-game multi can carry 20-40% vig — the bookmaker's margin compounds with each added leg. See the vig explainer for the full calculation mechanics.
A 4.5% negative expectation per bet sounds manageable. The problem is compounding. The probability of being in profit after N bets, assuming a 50% true win rate at $1.91 odds (the "coin-flip punter"), follows a binomial distribution. The expected value after N bets:
Expected return after N bets = N × (-0.045) × stake
After 100 bets at $100 each: expected loss = $450. After 500 bets: expected loss = $2,250. After 1,000 bets: expected loss = $4,500.
But expected value is just the centre of the distribution. The probability of actually being in profit after N bets — of random variance overcoming the negative expectation — drops rapidly:
This is the coin-flip punter — someone who is exactly as good as random chance at picking winners, betting at standard bookmaker odds. After 500 bets, they have a 92% probability of losing money. After 1,000 bets, 99%. This is not because they are bad at picking. It is because the vig is a mathematical certainty that variance can temporarily hide but cannot permanently overcome.
The coin-flip punter describes most recreational punters — people who watch the sport, have opinions about who will win, and bet those opinions at whatever price their preferred bookmaker is offering. The vig guarantees they lose over time. The only question is how fast.
To be profitable long-term, you need to find bets where the true probability is higher than the implied probability of the odds. At $1.91 odds, the implied probability is 52.36% (1/1.91). You need to win these bets at a rate above 52.36% to be profitable. At 53% win rate on $1.91 bets:
EV per bet = 0.53 × $0.91 + 0.47 × (-$1.00) = $0.0123 per dollar
A 1.23% edge. At 500 bets of $100 each: expected profit = $615. But even at a genuine 1.23% edge, the probability of being in profit after 100 bets is only about 55%. After 500 bets, about 61%. After 1,000 bets, about 65%. A genuine edge does not guarantee short-term winning — variance can drown a real edge for hundreds of bets. This is why so few punters persist long enough to find out whether they actually have an edge. The feedback loop is too slow and too noisy.
The table below shows the required win rate to break even at different odds levels:
The required win rate drops as the odds lengthen — but the variance increases dramatically. A $10.00 bet has a much lower breakeven win rate (10%) but you will lose 90% of the time and the losing streaks will be much longer. The bankroll swings are wilder. This is the trade-off at the core of sports betting: lower breakeven rates come with higher variance, and higher variance means you need a larger bankroll to survive the downswings. See the bankroll management guide for the framework.
Multi bets compound the vig across each leg. A 2-leg multi where each leg is priced at $1.91 (5% vig per leg):
Combined vig = 1 - (1/1.91 × 1/1.91) = 1 - 0.274 = ~9.3% edge to the bookmaker
A 3-leg multi of $1.91 legs: combined vig ≈ 13.6%. A 4-leg multi: ~17.7%. A 5-leg multi: ~21.5%. Same-game multis are worse because the legs are correlated and the bookmaker applies an additional correlation discount that is opaque to the punter. The $1.91 price on each leg of a same-game multi is not the same $1.91 you would get betting the leg individually — the bookmaker shortens the multi-leg price further to account for correlation risk.
This is the arithmetic behind the multi bet piece: the vig compounds geometrically with each added leg, and the punter's probability of winning compounds geometrically downward. A 5-leg multi of 50% true probability events has a 3.1% chance of winning. The bookmaker is pricing it as if each leg has a ~52.4% chance (the $1.91 implied probability). The gap between 3.1% and the bookmaker's implied probability is the multi vig. It is enormous.
The punters who finish in profit are not the ones with the best sport knowledge or the most sophisticated analysis. They are the ones who systematically find prices that are better than the true probability. The three main paths:
Arbitrage. Betting both sides of the same outcome at different bookmakers where the combined odds guarantee profit. No sport knowledge required — it is pure price mathematics. See the arbitrage guide.
Positive expected value betting. Identifying when a bookmaker's price is above the true probability (usually by comparing against the market consensus or exchange price). Requires some probability estimation skill but the edge comes from price comparison, not from being smarter than the bookmaker. See the +EV guide.
Closing line value. Betting at prices that are better than the eventual closing line. The closing line is the most efficient price — it incorporates all available information. Beating the closing line consistently is the strongest evidence of a genuine edge. See the CLV guide.
All three paths share the same core principle: the edge comes from the price, not from the prediction. The bookmaker's vulnerability is not that they misprice individual events — their models are good. The vulnerability is that different bookmakers price the same event differently, and those differences create pockets of positive expectation that can be systematically harvested. The winning punters are price comparison operators, not oracle predictors.
There is a behavioural reason most punters think they are doing better than they are: they do not track their results. Or they track selectively — remembering the big wins, forgetting the steady drip of small losses.
A 2024 survey of Australian punters found that fewer than 15% kept any form of betting record. Of those who did, fewer than half tracked net profit and loss — most tracked only win/loss count, which is meaningless without accounting for stake sizes and odds. A punter who wins 60% of their bets but bets $100 on favourites at $1.50 and $20 on underdogs at $5.00 is probably losing money — the win rate looks good but the dollars are negative.
The simplest edge most punters can give themselves: track every bet. Odds, stake, outcome, net result. After 500 tracked bets, the numbers will tell you whether you have an edge or whether the vig has been compounding against you. Most punters will discover the latter. The ones who discover the former are the ones who should keep betting. The rest should either stop or change what they are doing. See the bet tracker guide for a free setup.
Knowledge of a sport does not overcome the vig. If you bet Collingwood at $1.91 every week because you think they are a good team, and the true probability of Collingwood winning is 50%, you lose 4.5 cents per dollar bet regardless of how much AFL you watch. Knowledge only creates an edge if it allows you to identify prices that are wrong — cases where the bookmaker's implied probability is lower than the true probability. Most sport knowledge is already priced into the market. The bookmaker's AFL model incorporates team strength, home ground advantage, injuries, weather, and historical patterns. Your knowledge advantage over the market is probably smaller than you think. The vig is larger than you think. The math favours the vig.
No staking system turns a negative-expectation bet into a positive-expectation outcome. Staking systems change the distribution of outcomes — you win small most of the time and lose catastrophically occasionally — but they do not change the expected value. The EV of each individual bet is negative. Summing negative numbers does not produce a positive. A Martingale system (doubling after losses) will eventually hit either the table limit or the bankroll limit, and the one catastrophic loss will exceed all the accumulated small wins. The math is definitive on this point: staking systems do not work. Only finding positive-EV bets works.
James covers the AU bookmaker market — pricing mechanics, line movement, promotional structures, and how the corporate books actually operate. Previously worked in financial markets before moving to sports analytics.
NRL prop markets show wider cross-bookmaker price gaps in the second half of the season. Origin disruption, injury replacements, and slow model adjustments create real value for punters who know where to look.
