Every sports bettor wants to win. But most bettors approach it backwards: they pick a team they like, check the odds, and place the bet. Serious sports betting research doesn't work that way.

The professionals, the sharp bettors who actually make money long-term, do their bet research before they ever look at what they "like." They start with one question: does this bet have positive expected value?

If the answer is yes, they bet it. If not, they pass. No exceptions.

This guide explains exactly what expected value is, how to calculate it, and how to find +EV bets in a way you can repeat every single day.

What Is Expected Value in Sports Betting?

Expected value (EV) is the average amount you'd win (or lose) per bet if you placed the same bet thousands of times. It's expressed in dollars relative to a $100 wager.

A positive EV bet (+EV) means that, on average, you'd profit over time. A negative EV bet (-EV) means you'd lose money over time, even if you win individual bets.

The formula is straightforward:

Expected Value Formula

EV = (Win Probability x Profit) - (Loss Probability x Stake)

Let's walk through a simple example. Say you're betting on a coin flip (50/50) and a sportsbook offers you +110 odds (bet $100, win $110).

  • Win probability: 50% (0.50)
  • Profit if you win: $110
  • Loss probability: 50% (0.50)
  • Stake at risk: $100

EV = (0.50 x $110) - (0.50 x $100) = $55 - $50 = +$5

That's a +EV bet. For every $100 you wager, you'd expect to profit $5 on average over thousands of bets. You'll lose individual flips, but the math is on your side.

Now flip it: the same coin flip at -110 (bet $110, win $100).

EV = (0.50 x $100) - (0.50 x $110) = $50 - $55 = -$5

That's the vig. The sportsbook's cut. Almost every bet you see on a sportsbook has negative EV by default, because the books build their margin into the odds.

Key insight: Finding +EV bets means finding situations where the sportsbook has mispriced the odds. Their implied probability is lower than the true probability. That gap is your edge.

Step 1: Calculate True Probability

The hardest part of sports betting research is estimating the true probability of an outcome. The sportsbook's implied probability (what the odds suggest) is not the true probability. It's inflated by the vig.

To find the true probability, you need to do your own research. This means:

  • Injury reports: Is a key player out? Significantly changes win probabilities.
  • Recent form: How has each team performed in the last 5-10 games?
  • Matchup data: Historical head-to-head records, pace, defensive ratings.
  • Line movement: Which way has the line moved? Sharp money often moves lines in the direction of the true edge.
  • Model consensus: What do independent probability models say?

This is where most bettors give up. The research is tedious. It takes hours. And you need to do it for every bet, every day, to find the handful that actually have positive expected value.

Let Juice do the bet research for you.

Upload a screenshot of any bet, and Juice runs analysis through ChatGPT, Claude, and Gemini simultaneously to estimate true probability and calculate EV.

Download on the App Store

Step 2: Convert Odds to Implied Probability

Once you have the sportsbook's odds, you need to convert them to implied probability to compare against your true probability estimate.

For American odds:

  • Positive odds (e.g. +150): Implied probability = 100 / (odds + 100). At +150: 100/250 = 40%.
  • Negative odds (e.g. -200): Implied probability = |odds| / (|odds| + 100). At -200: 200/300 = 66.7%.

Use our free betting odds converter to do this instantly for any odds format.

But there's a catch: the implied probabilities for both sides of a bet add up to more than 100%. That's the vig. To get the true no-vig probability, you need to normalize. Our EV calculator handles this automatically.

Step 3: Find the Gap

Now you have two numbers:

  1. Your estimated true probability (from research)
  2. The sportsbook's implied probability (from the odds, after removing vig)

If your true probability is higher than the sportsbook's implied probability, you've found a +EV bet. The bigger the gap, the bigger the edge.

Example: You estimate a team has a 55% chance of winning. The sportsbook's no-vig implied probability is 50%. That 5% gap is positive expected value. Bet it.

Step 4: Compare Odds Across Sportsbooks

Here's a move that most casual bettors completely miss: line shopping.

Different sportsbooks offer different odds on the same game. The differences seem small, but they compound enormously over hundreds of bets. Consistently getting the best available line is one of the highest-ROI habits in sports betting research.

Example: You've identified a +EV bet on the Houston Rockets ML. FanDuel has them at +105. BetMGM has them at +115. DraftKings has them at +110. Bovada has them at +120.

If you always bet the best available line (+120 at Bovada), you're getting 15 extra cents per dollar compared to the bettor who defaults to FanDuel. Over 500 bets a year, that's a massive difference in profitability.

Juice compares odds across 18 sportsbooks including FanDuel, DraftKings, BetMGM, Caesars, Bet365, Bovada, Hard Rock Bet, BetRivers, Fanatics, theScore, MyBookie, BetUS, Fliff, PrizePicks, Underdog, Novig, ProphetX, and Kalshi, so you never leave value on the table.

Step 5: Size Your Bets with Kelly Criterion

Once you've found a +EV bet, the next question is how much to bet. Too little and you underutilize your edge. Too much and a bad run can wipe out your bankroll before the math catches up.

The Kelly Criterion is the mathematically optimal bet sizing formula:

Kelly Criterion

f = (bp - q) / b

where b = decimal odds - 1, p = win probability, q = loss probability

Most sharp bettors use a fraction of Kelly (typically 25-50%) to account for estimation error in their probability estimates. The goal is growth without ruin.

Practical tip: If you're not sure about your probability estimate, bet smaller. The Kelly formula penalizes overconfidence by recommending smaller bets when your edge is uncertain.

The Repeatable Sports Betting Research Workflow

Here's the full workflow, condensed into a process you can run on any bet:

  1. Identify the game. Focus on sports and markets you know well. NBA, NFL, and MLB offer the most liquid markets and the best data.
  2. Research the matchup. Injuries, form, pace, historical matchups. Look for information the market might be underweighting.
  3. Estimate true probability. Based on your research, what do you think the real win probability is? Use multiple sources and models.
  4. Compare to sportsbook lines. Convert the odds to implied probability. Is there a gap?
  5. Shop the line. Find the best available odds across all books you have access to.
  6. Calculate EV. Use our free sports betting EV calculator to confirm the bet is +EV.
  7. Size appropriately. Use Kelly or a fraction of Kelly based on your confidence.
  8. Bet and track. Record every bet. Track your actual win rate vs. your estimated probability. Over time this calibrates your model.

Why Most Bettors Lose

Most sports bettors lose not because they're bad at predicting games, but because they skip the bet research. They bet on instinct, take the first odds they see, and never compare lines across books.

The average recreational bettor is playing against a machine. Sportsbooks have entire teams of quants setting lines to maximize their take. The only way to compete is to do better research, faster.

That's exactly what Juice is built for. Upload a screenshot of any bet slip and get instant AI sports betting research from three independent models, complete with probability estimates, EV calculations, and odds comparison across 18 sportsbooks. The research that used to take hours now takes seconds.

Try Juice free for 3 days.

Screenshot any bet. Get AI analysis from ChatGPT, Claude, and Gemini. Find the best odds across 18 books. No spreadsheets required.

Download on the App Store

Frequently Asked Questions

How often do +EV bets actually win?

A +EV bet doesn't win every time. A bet with 55% true probability will lose 45% of the time. The edge shows up over hundreds of bets, not individual outcomes. Short-term results are noisy. Long-term, positive EV compounds into real profit.

Can I find +EV bets without a model?

Yes. Line shopping alone (always getting the best available odds) is a form of +EV betting. Focusing on markets with stale lines, betting into line movement from sharp books, and targeting markets with high information asymmetry (like player props on injury news) are all valid approaches.

What sports are best for finding +EV bets?

NBA and NFL offer the most data and liquidity, but also the most competition. College sports, lower-profile markets (second-half lines, quarters), and player props tend to have more mispricings because books spend less time on them.

How do I track whether my bet research is working?

Track every bet: the odds, your estimated probability, and the result. After 200+ bets, compare your actual win rate to your estimated win rate. If you're consistently estimating correctly, your EV estimates are valid. If you're consistently wrong in one direction, your model needs calibration.