How to calculate expected value in sports betting
Expected value, usually shortened to EV, is a way to estimate whether a bet is profitable over a large number of similar wagers. It does not predict the result of one match. A positive-EV bet can lose today, while a negative-EV bet can win, so the calculation is about long-term decision-making rather than certainty.
For Australian punters, EV is particularly useful because decimal odds make the maths straightforward. Whether you are betting on an AFL match in Melbourne, an NRL game in Brisbane, or a tennis tournament in Sydney, the same principles apply: estimate the true probability, compare it with the bookmaker’s price, and account for the stake.
What expected value means for a bet
The basic EV formula is:
EV = (probability of winning × profit if successful) − (probability of losing × amount lost)
With decimal odds, the net profit from a one-unit stake is the odds minus one. If a team is priced at 2.50, a $10 winning bet returns $25 in total, including the original stake, so the net profit is $15.
Suppose you estimate a team’s chance of winning at 45%, while a bookmaker offers odds of 2.50. The calculation for a $10 stake is:
EV = (0.45 × $15) − (0.55 × $10)
EV = $6.75 − $5.50 = $1.25
The expected profit is therefore $1.25 per $10 wager, or 12.5% of the stake. This does not mean you will make $1.25 on the next bet. It means that, if your probability estimate is accurate and you repeatedly find the same price, the average result should move towards that figure.
A negative EV indicates that the price does not justify the risk. A zero-EV bet is theoretically fair before transaction costs, although bookmakers generally build in a margin, meaning genuinely fair prices are uncommon.
Turning odds into implied probability
The quickest way to interpret decimal odds is to convert them into implied probability:
Implied probability = 1 ÷ decimal odds
Odds of 2.00 imply a 50% chance, 4.00 imply 25%, and 1.50 imply approximately 66.67%. This gives you the break-even probability before considering your own opinion.
For example, odds of 1.80 imply:
1 ÷ 1.80 = 0.5556, or 55.56%
You need to believe the selection wins more often than 55.56% for it to have positive expected value. If your assessed probability is 59%, the edge is present. If your estimate is 52%, the price is too short even if the team appears likely to win.
Markets with several outcomes require extra care. In a two-way tennis market, adding the implied probabilities of both players shows the bookmaker’s overround. If one player is 1.70 and the other is 2.20, the implied probabilities are 58.82% and 45.45%, adding to 104.27%. The extra 4.27% represents the bookmaker’s theoretical margin.
Australian operators may display prices through familiar brands such as TAB, while online bookmakers compete on markets and promotions. A boosted price can change the EV calculation, but bonus terms, minimum odds, and return limits also need to be considered before treating the boost as genuine value.
Building a realistic probability estimate
The difficult part is rarely the formula. It is deciding whether your probability is credible. Start with information that has a logical relationship with the market: injuries, expected line-ups, venue, travel, weather, rest, surface, recent performance, and tactical match-ups.
For Australian sports, local conditions can matter. A wet night at the MCG may affect an AFL total-points market differently from a dry afternoon at Optus Stadium. In the NRL, travel from Townsville or Auckland can influence preparation and fatigue, while a fast hard court at the Australian Open can shape tennis pricing. These factors should adjust a baseline estimate rather than replace it.
Avoid treating recent results as proof of future probability. A team may have won three straight matches against weak opposition, or a striker may have scored several goals from an unusually high number of chances. Separating sustainable performance from short-term variance is essential.
The Kelly Criterion can help connect probability and staking, but it should come after the EV assessment. For broader probability education, practical learning resources can also help you improve your statistical foundations. Many bettors use a fractional Kelly approach because full Kelly staking can create uncomfortable swings when estimates are uncertain.
| Decimal odds |
Break-even probability |
Probability needed for positive EV |
| 1.50 |
66.67% |
Above 66.67% |
| 1.80 |
55.56% |
Above 55.56% |
| 2.00 |
50.00% |
Above 50.00% |
| 2.50 |
40.00% |
Above 40.00% |
| 3.00 |
33.33% |
Above 33.33% |
| 5.00 |
20.00% |
Above 20.00% |
Comparing price, edge and expected return
A useful distinction is the difference between probability edge and value edge. If you estimate a selection at 55% and the odds imply 50%, your probability edge is five percentage points. The EV calculation translates that difference into an expected financial return.
For a one-unit stake, a compact formula is:
EV = (your probability × decimal odds) − 1
Using a 55% estimate at 2.00 odds:
EV = (0.55 × 2.00) − 1 = 0.10
That is a 10% expected return per unit staked. A $50 bet would have an expected profit of $5, provided the estimate is sound. This is an average over many comparable bets, not a promised payout.
Shop around before placing a wager. A price of 2.00 may be available at one bookmaker while another offers 1.90. At an estimated 55% probability, odds of 2.00 create a 10% EV, but odds of 1.90 create:
(0.55 × 1.90) − 1 = 0.045
The expected return falls to 4.5%. That difference compounds over time, especially for regular punters betting the same sport.
Promotional bets need separate treatment. A $50 bonus bet may not return the stake, and a cash-out offer can settle below the mathematical value of holding the position. Read the Australian operator’s terms, including wagering requirements, expiry dates and eligible markets, before including a promotion in your calculations.
Managing uncertainty and recording results
Your probability is an estimate, not an objective fact. It may be based on a model, a betting exchange price, a ratings system, or personal analysis. Every method has errors, so avoid assigning extreme confidence without strong evidence. A selection assessed at 70% is still expected to lose three times in ten under the model.
Keep a record of the odds taken, your estimated probability, stake, result, closing price and reason for the bet. The closing line can be a useful benchmark: if your 2.10 price regularly shortens to 1.90 before the event, your timing may be strong even when individual results are mixed.
Variance is unavoidable. A good AFL multi can fail because of one late upset, and a Melbourne Cup outsider can win without proving that every long-shot bet was wise. Judge performance over a meaningful sample rather than reacting to one weekend of footy or a single winning streak.
Responsible betting also belongs in the calculation. Set a fixed bankroll, use modest stakes, and never increase the wager to recover losses. The maths only has value when your estimates, records and staking habits remain disciplined.
Use EV as a filter before every wager: convert the odds, form a defensible probability estimate, calculate the expected return, and compare prices across licensed Australian bookmakers. Readers wanting to strengthen their broader decision-making can also review Texas Hold’em starting hands because poker teaches the same core habits of probability, position and disciplined selection. Keep your records, protect your bankroll, and place a bet only when the price offers a reasoned edge rather than a passing hunch.