What Is Expected Value in Sports Betting?

Expected value (EV) is the average outcome you can expect from a bet if you repeated it an infinite number of times under identical conditions. It is not a prediction of what will happen on any single wager. It is a long-run mathematical expectation that tells you whether a bet is profitable or unprofitable at the fundamental level before variance clouds the picture.

Start with a coin flip. A fair coin lands heads 50% of the time. If someone offers you $1.10 for every $1.00 you bet on heads, your expected value per bet is positive because the payout exceeds the true probability-adjusted cost. If they only pay you $0.90 per dollar on a 50/50 flip, your EV is negative. Nothing changes about the coin. The edge lives entirely in the relationship between the true probability and the price offered.

Sports betting works the same way. Every bet you place has an implied probability baked into the odds and a true probability that actually reflects the likelihood of the outcome. Your job as a sharp bettor is to identify gaps between those two numbers. When the true probability of an outcome is higher than what the odds imply, you have a positive expected value bet. When the true probability is lower, you have a negative expected value bet.

This distinction matters more than whether your last ten bets won or lost. A bettor who placed ten straight winners on -EV bets got lucky. A bettor who went 4-6 on genuinely +EV bets is executing correctly. The long run is the only scoreboard that matters, and the long run always reflects expected value.

Understanding EV is not optional for bettors who want to win long-term. It is the single framework that separates disciplined, process-driven bettors from gamblers chasing results. Learn to how implied probability and vig work inside every betting line and you will have the foundation to calculate EV on any wager you consider.

The EV Formula: Breaking Down Every Component

The expected value formula for a sports bet is straightforward once you break it into its components. Here it is in full:

EV = (P_win x Profit) – (P_lose x Stake)

Each variable does specific work. P_win is your estimated true probability that the bet wins, expressed as a decimal between 0 and 1. Profit is the net amount you receive if the bet wins, meaning the payout minus your original stake. P_lose is your estimated true probability that the bet loses, which is simply 1 minus P_win. Stake is the amount you are risking on the bet. Every single sports bet you ever place can be evaluated through this four-variable lens.

Before you can plug numbers into the formula, you need to convert American odds into implied probability. For positive odds (underdog lines), the formula is: Implied Probability = 100 divided by (Odds + 100). For negative odds (favorite lines), the formula is: Implied Probability = the absolute value of Odds, divided by (the absolute value of Odds plus 100).

Take a -110 line, which is the most common odds price you will see on spread bets. Plug it in: 110 divided by (110 + 100) equals 110 divided by 210, which equals 0.524, or 52.4%. That is the sportsbook’s implied probability for that side. But notice: if both sides of the spread are priced at -110, the implied probabilities add up to 104.8%, not 100%. That extra 4.8% is the vig. The book’s true break-even probability on each side, stripped of juice, is 50%. The gap between 52.4% and 50% is the built-in house edge.

Now apply the EV formula to that -110 spread bet, assuming you correctly estimate the true probability at 50% (the no-vig fair value). With a $110 stake, your profit if you win is $100.

The EV on a -110 bet where you only have 50% true probability is negative $5.00 per $110 wagered. That is a -4.55% return on every dollar you stake. Multiply that across hundreds of bets per season and you see precisely why the house wins over time. To break even at -110, you need to win at least 52.4% of those bets. To generate positive EV, you need to win more than 52.4% consistently, which requires your true probability estimate to exceed the book’s implied probability after vig removal.

How to Calculate EV on a Real Bet: Step-by-Step

The formula only pays off when you apply it systematically to real betting decisions. Walk through this five-step process on every bet you seriously consider and you will start thinking in expected value rather than gut feel.

1. 01

   Identify the Posted Odds

   Find the odds at your best available sportsbook for the bet you are considering. Write them down. For this example, you are looking at an NFL moneyline: an underdog priced at +150.

2. 02

   Convert Odds to Implied Probability

   Use the formula for positive odds: 100 divided by (150 + 100) equals 100 divided by 250 equals 0.40, or 40%. That is the sportsbook’s implied probability that this team wins. Remember, this number still contains the vig, so the fair-value probability the book actually believes in is slightly higher than 40%.

3. 03

   Estimate Your Own True Probability

   This is the step that separates sharp bettors from everyone else. Using your model, power ratings, injury reports, situational research, and any other relevant information, you estimate this team’s actual probability of winning at 45%. You believe the book has underpriced this team by 5 percentage points.

4. 04

   Plug Both Numbers Into the EV Formula

   Your stake is $100. If you win at +150, your profit is $150. Your P_win is 0.45 and your P_lose is 0.55. Calculate: EV equals (0.45 x $150) minus (0.55 x $100). That equals $67.50 minus $55.00, which equals positive $12.50.

5. 05

   Interpret the Result in Dollar Terms

   Your EV is plus $12.50 per $100 staked, or a 12.5% edge on this bet. Over 100 identical bets at this edge, you would expect to profit approximately $1,250. This does not mean you win every bet. It means the math is working in your favor on every single wager you place at this price.

Let’s see the full arithmetic laid out cleanly so you can replicate it yourself on any bet.

Breaking it down step by step: odds are +150, implied probability from book equals 40%, your true probability estimate equals 45%, profit if win equals $150, stake equals $100, P_lose equals 0.55. The calculation is (0.45 x 150) minus (0.55 x 100), which gives you 67.50 minus 55.00 equals +12.50.

This +12.50 figure is the expected value per bet. Expressed as a percentage of your stake, it is 12.5% EV. That is an unusually large edge by professional standards, which suggests either the book has made a significant mispricing or, more likely, you need to challenge your own 45% estimate. Real sharp edges on NFL moneylines typically run 2 to 5 percentage points, not 12. The math is clean; the hard work is in the probability estimate.

One practical note on bet sizing: the EV calculation tells you whether to bet, but it does not automatically tell you how much. Your unit size should always be calibrated to your overall bankroll and your confidence level in the edge. Pairing this EV process with a disciplined Kelly Criterion sizing to maximize long-run EV growth framework ensures you do not blow out your bankroll during the inevitable losing stretches that variance produces even on +EV plays.

Positive EV vs. Negative EV Bets: What the Difference Really Means

A positive EV bet is one where your true probability of winning exceeds the probability implied by the odds after vig removal. A negative EV bet is the opposite: the odds imply a higher probability than the true chance of winning. In dollar terms, +EV means you expect to profit over a large sample. Negative EV means you expect to lose money, slowly or quickly depending on the size of the edge against you.

Here is the critical point that recreational bettors miss: a +EV bet can absolutely lose, and a -EV bet can absolutely win. Expected value is a statement about averages across hundreds or thousands of bets, not about any single outcome. If you bet a +8% EV play and lose, you made the right decision. If you bet a -8% EV play and win, you got lucky, but you made the wrong decision. Process and outcomes are separate things in the short run.

The table below shows how EV changes as your true probability estimate shifts relative to different posted odds. Notice how the breakeven point and edge percentage move together.

The long-run edge percentage is the number that defines your operation as a sharp bettor. It is your average EV per bet expressed as a percentage of your stake. Most professional bettors operate with edges between 2 and 5% per bet. Those numbers sound small, but at volume they compound into meaningful profit. A bettor placing 500 bets per year at an average $200 stake with a 3% edge generates $3,000 in expected profit annually before accounting for variance. The math rewards consistency and patience more than it rewards any single big score.

How to Estimate True Probability: The Key Input Sharps Get Right

The EV formula is only as good as the probability estimate you plug into it. This is the part of the process that separates bettors who beat the market from those who just think they do. Get the probability estimate wrong by a few percentage points and you will confidently bet negative EV plays all season, wondering why the math is not working.

The most reliable starting point for estimating true probability is the no-vig closing line at a sharp book. Sharp sportsbooks like Pinnacle, Circa, and Bookmaker process enormous volume from professional bettors and have strong incentives to price lines accurately. By the time a game starts, their closing lines reflect the collective information of the sharpest money in the market. Strip the vig from those closing odds and you have a close approximation of the market’s consensus true probability.

Beyond market prices, the sharpest bettors build their own probability estimates through several methods. Power ratings assign numerical strength values to each team and translate point-spread differences into win probabilities. Regression-based models use historical data to weight factors like yards per play, turnover margin, offensive and defensive efficiency, and pace. Situational models layer in non-statistical factors: rest advantages, divisional familiarity, travel schedules, and weather for outdoor games.

No single method is complete. The best approach combines your own model output with what the sharp market is telling you. If your model says a team should be favored by 4 points but the sharp closing line is -7, the market likely knows something your model does not. Investigate before fading the market blindly.

A 2 to 3% edge per bet is meaningful in practice. At 500 bets per year with a $250 average stake and a 2.5% edge, that is $3,125 in expected annual profit. The same edge at $1,000 per bet generates $12,500. The edge itself does not need to be large to generate real income. It needs to be real, persistent, and based on accurate probability estimation. The playbook on identifying value bets before the market corrects covers specific research methods for building those probability estimates across different sports and bet types.

One practical tool worth using is an EV calculator or no-vig probability converter. Enter the sportsbook’s odds and your own estimated probability, and the tool outputs the EV percentage immediately. This removes arithmetic errors from the process and lets you evaluate more bets faster. Speed matters in markets that move quickly after opening.

Applying EV Calculation to Different Bet Types

The EV formula does not change based on what type of bet you are making. What changes is how you estimate the true probability for each market type. Moneylines, spreads, totals, and player props each require a different research approach, but the final output always feeds into the same equation: EV equals (P_win x Profit) minus (P_lose x Stake).

On moneylines, your probability estimate is a direct win probability for one team. On point spreads, you are estimating the probability that one team covers a given margin, which requires both a point projection and an understanding of how likely that specific margin difference is. On totals, you are estimating the probability the combined score falls above or below the posted number, which requires offensive and defensive output projections for both teams.

Player props deserve special attention. Books often post props on secondary markets with less precision than they apply to game lines, particularly early in the week. Usage rate, matchup data, pace of play, and injury context are all inputs that sharp bettors can exploit when the book’s number reflects outdated information. For a detailed breakdown of that research process, see the full guide on applying EV analysis to NBA player prop markets.

Parlays deserve a direct warning. The EV formula applied to a two-leg parlay requires multiplying the probabilities of both legs, and since each leg already carries negative EV from the vig, the combined EV is more negative than either leg alone. A two-team parlay at standard juice carries roughly double the vig burden of a single bet. Three-team parlays and beyond accelerate that drag. Sharp bettors generally avoid parlays unless specific conditions apply, such as correlated outcomes where a team winning the game also tends to cover the spread, or promotional odds boosts that shift the math in your favor.

What Is a Good +EV Percentage for Sports Bettors?

This is one of the most commonly asked questions from bettors who are starting to apply EV thinking to their process. The answer requires context, but there are clear professional benchmarks that give you a realistic target to work toward.

Professional sports bettors typically operate on edges of 2 to 5% per bet. That range might sound underwhelming, but it is sustainable, repeatable, and extraordinarily profitable at meaningful stake sizes and volume. Anything above 5% per bet maintained consistently is exceptional and almost always reflects either unusual market access, proprietary information, or a concentrated niche where a bettor has developed a genuine information advantage over the book.

Sample size is the context that most bettors ignore. You cannot conclude that you have a 4% edge from 50 bets. Variance at those sample sizes is large enough to produce a 4% ROI through pure luck. You need at least 500 bets at similar odds before your results start converging toward your true edge. Under 200 bets, the short-run win rate is almost entirely dominated by variance rather than skill.

Here are the ROI benchmarks that define each tier of bettor performance over large samples:

Understanding where you fall in this spectrum requires honest record-keeping and patience. Most bettors overestimate their edge because they recall winning bets more vividly than losses. The only defense against that bias is a complete, unedited betting log that includes every single wager you placed, including the ones you would rather forget.

EV and Closing Line Value: Why CLV Is the Best Proof of Edge

Closing line value, or CLV, is the single most powerful proxy for expected value available to sports bettors. It answers a specific question: did you get a better price than the market settled on by game time? If you consistently beat the closing line across hundreds of bets, you are consistently buying at below-fair-value prices, which is the definition of positive expected value betting.

Here is how it works in practice. You bet a team at -105 on Monday morning. By kickoff on Sunday, the same team is priced at -115. You beat the closing line by 10 cents. That means the market moved in the direction of your opinion, validating that your bet was made at a price better than the market’s final consensus. Over a large sample, bettors who consistently beat the closing line are demonstrating real edge regardless of what their short-run win rate looks like.

Sharp sportsbooks like Pinnacle are the gold standard for closing line benchmarks. Pinnacle accepts large bets from professional bettors and does not limit winners, so their closing prices reflect maximum available information. When you beat Pinnacle’s closing line, you are beating a market that sharp money has already sharpened. That is meaningful validation of your process.

The relationship between CLV and EV is not coincidental. They are measuring the same underlying reality from different angles. EV looks forward from your estimated true probability. CLV looks backward from the market’s final verdict. When both consistently point in the same direction, you have a betting process worth scaling. Understanding the mechanics behind why odds move, who moves them, and why sharp books set efficient closing lines starts with understanding the full picture of how implied probability and vig work inside every betting line.

Common EV Mistakes Bettors Make (And How to Avoid Them)

Most bettors who understand the EV concept intellectually still make errors that undermine its application. The mistakes below are not beginner oversights. They are process failures that affect intermediate and advanced bettors regularly.

1. 01

   Using the Book’s Implied Probability as Your True Probability

   The most common error. If you use the sportsbook’s implied probability (which includes vig) as your P_win estimate, your EV calculation will always show near-zero or negative values because you are measuring the line against itself. Always strip the vig first and then compare your independent estimate to the no-vig fair value.

2. 02

   Ignoring Vig When Estimating Your Edge

   Bettors sometimes calculate their edge as the difference between their probability estimate and the raw implied probability without removing the house cut. This overstates the edge by 2 to 5 percentage points depending on the market. Always compare your true probability to the no-vig market probability, not the listed implied probability.

3. 03

   Drawing Conclusions From Small Samples

   If you are 12-4 over your last 16 bets and conclude you have a 75% win rate edge, you are making a statistically meaningless inference. You need 500 or more bets at consistent odds before short-run results tell you anything reliable about true edge.

4. 04

   Overestimating Edge Due to Confirmation Bias

   Bettors remember the research that supported their winning bets and forget the research that led to losing ones. The result is a systematically inflated self-assessed edge. Log every bet with your reasoning beforehand, not after the result is known. That paper trail keeps you honest.

5. 05

   Betting Familiar Markets With No Edge

   Many bettors default to NFL spreads or the NBA because they watch those games constantly. Familiarity is not an edge. If the market is efficient in a given sport or bet type and your model has no information advantage, your EV is negative by the amount of the vig. Be willing to bet less-familiar markets where the book is thinner.

These process failures compound quietly over time. The bettor who avoids all five of them does not necessarily win more individual bets. They win more of the right bets at the right prices, and that is the only distinction that matters across a full season. For a broader inventory of the process errors that cost bettors the most money, the full breakdown of common betting mistakes that destroy your long-run EV covers dozens of specific scenarios with concrete fixes.

Building a Repeatable +EV Betting Process: Your Action Plan

Everything covered in this tutorial converges on one practical question: how do you build a betting operation that generates consistent positive EV week after week? The answer is a system, not a feeling. Here is how to build one from the ground up.

1. 01

   Build Your Probability Framework

   Choose one or two sports to start. Develop a simple power rating system for each team, updated weekly after results. Use efficiency metrics (yards per play for NFL, offensive and defensive rating for NBA) rather than raw scores. Your goal is a probability estimate for every game you consider betting that is independent of the sportsbook’s price.

2. 02

   Shop Lines Across Multiple Sportsbooks

   Open accounts at a minimum of four sportsbooks, including at least one sharp book like Pinnacle for reference. Before placing any bet, check all available prices. Getting +155 instead of +148 on a moneyline raises your EV by several percentage points. Line shopping is free money and bettors who skip it are voluntarily reducing their edge.

3. 03

   Log Every Bet With Your Pre-Game Estimate

   Record the posted odds, your estimated true probability, your calculated EV percentage, and your unit size for every single bet before the game starts. After the game, record the closing line. This log is your performance data and your protection against confirmation bias.

4. 04

   Apply Disciplined Staking Based on EV and Kelly Sizing

   Use flat staking or a fractional Kelly approach to size bets relative to your estimated edge and bankroll. Betting larger on higher-EV spots and smaller on marginal edges is how you maximize long-run growth. The full mechanics of this approach are covered in the Kelly Criterion sizing to maximize long-run EV growth framework.

5. 05

   Review Weekly and Adjust

   Every week, review your average CLV, your EV per bet, and your actual ROI. Look for markets where your CLV is consistently positive (your model is identifying value) and markets where it is consistently negative (your model is lagging the market). Double down on your strengths. Cut markets where you have no demonstrable edge.

The bettors who win long-term are not the ones who know the most about sports. They are the ones who have the most disciplined process for converting sports knowledge into accurate probability estimates and then finding prices that exceed those estimates. Build the system, trust the sample, and let the expected value do the work.