Use a simple process to compare implied probability to your projection and identify positive expected value wagers.
New to this? Start with:Implied Probability Guide
The Cowboys are listed at +150 tonight. Your buddy says "take it, they're due for a win." But you have a model that says Dallas wins this game 42% of the time, and the odds imply only 40%. That 2-point gap is not a gut feeling — it is a value bet.
Value betting is not about picking winners. It is about finding mispriced odds. A bet has value when the sportsbook offers a price that underestimates the true likelihood of an outcome. If you can consistently identify those gaps, you will profit over time regardless of any single result. This guide walks through the complete process, from converting odds to sizing your bets.
A value bet exists whenever the probability you assign to an outcome is higher than the probability implied by the odds.
Think of it like buying a dollar coin for 90 cents. The coin is worth $1 every time, but you only paid $0.90. That 10-cent gap is your edge. In betting, the "coin" is the true probability of winning, and the "price" is what the sportsbook charges you through the odds.
The sportsbook sets odds based on their models, market demand, and the margin they build into each line. When their assessment differs from the true probability, a gap opens. That gap is the edge that value bettors exploit.
Here is the core principle: you do not need to be right more often than wrong. You need to be right more often than the odds imply. A team that wins 40% of the time can still be a value bet if the odds imply only a 33% chance.
The original three-step version of this process works, but serious value betting requires five steps. Each one builds on the last.
Before you can assess value, you need to know what the sportsbook thinks the probability is. That means converting their odds into a percentage.
American odds formulas:
Decimal odds formula:
For example, American odds of +150 convert to: 100 / (150 + 100) = 100 / 250 = 0.40, or 40%.
Keep in mind that sportsbook odds include a built-in margin (the vig or juice). For a cleaner picture of the market's true view, remove the vig first using a No-Vig Calculator. For a deeper explanation of implied probability and its formulas, see the Implied Probability Guide.
Use the Odds Converter to handle the math instantly across American, decimal, and fractional formats.
This is the hardest step and the one that separates profitable bettors from everyone else. You need your own honest assessment of how likely an outcome is.
Common methods include:
No method is perfect. The goal is not certainty. The goal is to be more accurate than the sportsbook's implied probability, on average, over time.
Once you have both numbers, the math is simple.
Edge = Your Estimated Probability - Implied Probability
For example, if you estimate a team at 42% and the implied probability is 40%, your edge is +2 percentage points. That 2% is the theoretical advantage you hold on this bet.
Not every positive edge is worth betting. The edge must be large enough to overcome variance and any potential error in your model. Most experienced value bettors look for a minimum edge of 2-3% before acting.
Edge tells you the direction. Expected value tells you the magnitude in dollars and cents.
EV = (Your Probability x Profit if Win) - (Lose Probability x Stake)
Using the example above with a $100 bet at +150 odds and your 42% estimate:
A positive EV means the bet is profitable in expectation. Over many repetitions, you expect to earn $5 for every $100 you put at risk on bets like this.
For a full breakdown of the formula and its components, see the Positive EV Betting Formula Guide. Use the EV Calculator to run the numbers quickly.
Finding value is only half the equation. You also need to bet the right amount.
Kelly Criterion is the most well-known approach. It sizes your bet proportionally to your edge: larger edge means a larger bet, smaller edge means a smaller bet. The formula is:
In practice, the full Kelly recommendation is too aggressive for most bettors because it assumes your probability estimate is perfectly accurate. Most professionals use fractional Kelly (half Kelly or quarter Kelly) to reduce the risk of large drawdowns.
Flat betting is the simpler alternative. You wager the same fixed amount (typically 1-3% of your bankroll) on every qualifying bet. It sacrifices some theoretical efficiency but is easier to manage and more forgiving of estimation errors.
Either way, never increase your stake based on gut feeling or recent results. Let the math determine the size.
Suppose the Dallas Cowboys are listed at +150 (moneyline) against the Philadelphia Eagles.
Step 1: Convert to implied probability.
100 / (150 + 100) = 40.0%
The sportsbook prices Dallas at a 40% chance to win.
Step 2: Estimate your probability.
Your model, which accounts for injuries, rest days, and historical matchup data, outputs Dallas at 42%.
Step 3: Calculate the edge.
42% - 40% = +2.0% edge
Step 4: Calculate expected value.
On a $100 bet:
Step 5: Size the bet.
Using a half-Kelly approach with a $5,000 bankroll, the recommended stake would be approximately 1% of bankroll, or $50. A flat bettor staking 2% would wager $100.
Verdict: This is a value bet. The edge is positive and the expected value confirms it. Place the bet at the appropriate size.
A sportsbook offers Over 45.5 total points in an NFL game at -110 (American odds).
Step 1: Convert to implied probability.
110 / (110 + 100) = 52.4%
Step 2: Estimate your probability.
Your analysis suggests the over hits about 50% of the time.
Step 3: Calculate the edge.
50% - 52.4% = -2.4% edge
The edge is negative. The sportsbook is pricing this outcome higher than your estimate. There is no value here.
Verdict: Pass. No matter how much you like the over, the math says this price is not in your favor. Discipline means walking away from bets without value, even when your instinct disagrees.
Value betting is a volume game. A small edge compounds over many bets, but the short term is noisy.
Here is what a consistent +3% edge looks like over time, assuming flat $100 bets at average odds of -110:
| Bets placed | Expected profit | Typical range (1 standard deviation) |
|---|---|---|
| 100 | +$300 | -$200 to +$800 |
| 500 | +$1,500 | +$250 to +$2,750 |
| 1,000 | +$3,000 | +$1,250 to +$4,750 |
At 100 bets, it is entirely possible to be losing money even with a real edge. At 1,000 bets, the probability of being in profit is very high. This is why sample size matters more than any individual result.
Variance is the reason bankroll management exists. Even with a genuine edge, losing streaks of 10, 15, or 20 bets in a row will happen. Your bankroll must survive those stretches for the math to play out.
The most dangerous mistake. If you think your edge is 5% but it is actually 1%, you will bet too aggressively and face larger drawdowns than expected. Always be conservative with your probability estimates. When in doubt, shrink your estimated edge.
Odds change for a reason. If you identify value at +150 but the line moves to +130 before you bet, recalculate. The value may have disappeared. Conversely, if the line moves in your favor (to +160), the value has increased.
Judging your strategy after 50 bets is like flipping a coin 10 times and concluding it is unfair. You need hundreds of bets at minimum to draw meaningful conclusions. Track everything, but resist the urge to overhaul your approach after a small number of results.
Value betting is mechanical. If your model says pass, you pass. Betting on your favorite team because "it feels right" or chasing losses with larger bets are the fastest ways to destroy your bankroll. Emotion is the opposite of edge.
Always account for the sportsbook's margin. A bet at -115 has worse implied probability than the same bet at -105. Shopping lines across multiple books to find the best price is one of the simplest ways to increase your effective edge.
Tip: Start with one sport you know well. Deep domain knowledge is a legitimate edge over sportsbooks that set lines across dozens of sports. You do not need a supercomputer — you need a systematic process and the discipline to track your results.
You do not need a supercomputer. You need a systematic approach.
Start with markets you know. If you follow the NFL closely, start there. Deep domain knowledge is a legitimate edge over sportsbooks that must set lines across dozens of sports.
Track your estimates versus results. For every bet, record your estimated probability, the implied probability, and the outcome. After 200+ bets, compare your estimates to your actual win rate. If they diverge significantly, adjust your model.
Use closing lines as a benchmark. The closing line at sharp sportsbooks is the market's final, most efficient probability estimate. If your pre-game estimates consistently beat the closing line, your model has real predictive power. If they do not, your estimates need work.
Be honest about what you do not know. No model accounts for everything. Injuries reported minutes before tip-off, weather changes, and lineup surprises are real factors. Build uncertainty into your estimates rather than pretending you have perfect information.
Use these calculators to run value betting math without errors:
A value bet is when your estimated win probability is higher than the implied probability from the odds. The difference is your edge.
Yes. Value is a long-term concept. Individual bets can lose even if they have positive expected value. The edge only shows up over hundreds or thousands of bets.
Subtract the implied probability from your estimated probability. For example, if you estimate 42% and the implied probability is 40%, your edge is 2 percentage points.
The larger your sample, the more reliably your results will reflect your edge. Most serious value bettors track at least 500 to 1,000 bets before drawing conclusions.
The Kelly Criterion is a formula that determines optimal bet size based on your edge and the odds offered. Many bettors use a fractional Kelly (such as half or quarter Kelly) to reduce variance.
No. Arbitrage locks in a guaranteed profit by betting both sides across different books. Value betting bets one side where you believe the odds are mispriced, accepting short-term variance for long-term profit.
Track your predictions against actual outcomes over time. Compare your estimates to closing lines at sharp sportsbooks. If your projections consistently beat closing odds, your model has genuine skill.