Consider a coin flip priced at +110 (win $110 on heads, lose $100 on tails). If the coin is fair (50% probability), the expected value is (0.5 × 110) - (0.5 × 100) = $5 per flip. That means, on average, you profit $5 per flip over millions of flips. In reality, variance means you might lose the first ten flips. But the math does not change. The edge is real, and the results will converge to the expected value over a large enough sample.
The problem for most bettors is that they never calculate EV. They bet on teams they like, games they watch, or lines that look attractive without converting the price into a probability and comparing it to their own estimate. A -150 favorite looks safe, but if the true probability is 55% and the implied probability at -150 is 60%, you are paying five cents per dollar for a coin that is slightly worse than fair.
Practical EV calculation starts with converting the sportsbook odds to implied probability, then adjusting for vig. Next, estimate the true probability using your own analysis, historical data, or a model. If your estimate exceeds the no-vig implied probability by a meaningful margin, the bet has positive EV. The margin you require should increase with your uncertainty. A bet you are less confident about should need a larger edge to justify the risk.
EV is not about winning individual bets. It is about the long-term trajectory of your bankroll. A bettor who consistently places positive EV bets will profit over time, even if they lose more individual bets than they win. A bettor who places negative EV bets will lose over time, even if they win more often than they lose. The scoreboard is the bankroll, not the win-loss record.