What is Positive EV Trading? Complete Beginner Guide (2026)

Expected value (EV) is the single most important concept in any form of betting, trading, or gambling. Understanding it is the difference between people who consistently make money and people who think they're "due for a win." This guide explains positive EV trading from scratch, with concrete examples, and covers when it works and when it doesn't.

## The One-Sentence Definition

Positive EV trading means making bets where the long-run expected profit is greater than zero, regardless of any individual outcome.

That last phrase matters. A +EV trade can still lose. In fact, plenty of them do. What makes it +EV is that if you could repeat the same situation thousands of times, your average outcome would be positive.

## The Math (It's Simpler Than You Think)

Expected value is just probability times payoff, summed across all possible outcomes.

For a coin flip where you win $2 if heads, lose $1 if tails: - EV = (0.5 × $2) + (0.5 × -$1) = $1.00 − $0.50 = +$0.50

If you could play this game infinitely many times, you'd average 50 cents profit per flip. The first 10 flips might lose you $4. The next 10 might win you $14. Over 1000 flips, you'd be up roughly $500. The variance disappears with sample size.

That's the math. Now the harder part.

## The Catch: Your Probability Has to Be Right

The formula EV = (p × payoff) − ((1-p) × cost) only gives you positive EV if your probability p is accurate. If your estimate is wrong, your EV is wrong.

Example: a Polymarket contract trades at $0.40 for YES. You think the true probability is 55%. Math says: - EV per share = (0.55 × $0.60) − (0.45 × $0.40) = $0.33 − $0.18 = +$0.15 per share

15-cent profit per share looks great. But what if the true probability is actually 42%, not 55%? Then: - EV per share = (0.42 × $0.60) − (0.58 × $0.40) = $0.252 − $0.232 = +$0.02 per share

Still positive, but barely — and well within the margin where bad luck on a few trades wipes out years of profit.

The lesson: +EV depends entirely on your probability estimate being more accurate than the market's. If you're wrong, you're not making +EV trades, you're just making bets you happen to like.

## Why +EV Trading Works in Prediction Markets

In efficient markets like the S&P 500, the consensus price reflects everything everyone knows. Beating it requires either insider information (illegal) or being smarter than thousands of professional traders (rare).

Prediction markets are different. They're newer, smaller, less liquid, and dominated by enthusiast retail traders. The "wisdom of crowds" works in aggregate, but specific markets — especially in niche topics — often have systematic mispricings.

The most common biases:

**Favorite-longshot bias**: traders overpay for unlikely outcomes ("$5 on this 90% market won't hurt") and underprice favorites. A 90% market often resolves YES 92-94% of the time historically — small but real edge.

**Recency bias**: latest news gets weighted too heavily, base rates get ignored. After a single bad poll, an election market crashes more than the data warrants.

**Narrative-driven pricing**: dramatic stories pull prices toward the "good story" outcome. The boring outcome is often underpriced.

**Sample-size errors**: traders confidently price unusual events based on tiny historical samples. "This hasn't happened in 10 years" doesn't tell you the probability of it happening now.

A trader who systematically identifies these biases — and trades against them with disciplined sizing — captures +EV over time.

## When +EV Trading Fails

Just because the framework is right doesn't mean execution is easy. Common failure modes:

### 1. Overconfidence in your estimates

You read three news articles and convinced yourself a market is mispriced. Maybe it is. But you're not factoring in: - The market participants who also read those articles - The participants who read different sources - Information you don't have access to

Rule of thumb: if your edge feels obvious, it's probably wrong. Real edges are usually 3-8 percentage points, not 30.

### 2. Position sizing errors

For the full framework, see our [Kelly Criterion guide](/blog/kelly-criterion-position-sizing).

Kelly Criterion says: bet a fraction of your bankroll proportional to your edge divided by variance. For typical prediction market edges, that's 1-5% of bankroll per trade.

Most retail traders either bet too small (no impact) or too big (one bad trade ruins them). Both kill long-run +EV.

### 3. Variance underestimation

A 60% edge bet means you LOSE 40% of the time. If you make 10 trades, expect to lose 3-5 of them. If losses freak you out and you stop following your process, your "+EV" strategy becomes -EV in practice.

You need enough bankroll to survive 10-15 consecutive losses on different bets. If you don't, you'll go broke before the math works in your favor.

### 4. Costs eating returns

For minimizing gas costs specifically, see our [Builder Program guide](/blog/gas-fees-builder-program-polymarket).

Polymarket has zero trading fees but gas costs ($0.10-2 per trade without Builder Program). On small positions, gas can be 5-10% of your trade size — easily wiping out a 3-5pp edge.

Kalshi has small fees (3-7% of profit) that similarly eat thin edges.

You need either large enough position sizes or low enough cost structure for +EV math to work after costs.

## Where to Find +EV in Prediction Markets

For the systematic scanning workflow, see our [how to find +EV markets guide](/blog/how-to-find-ev-markets-polymarket).

Markets where you have an information edge: - Domain you know professionally (your industry, your country's politics, your sport) - Niche topics that don't attract sophisticated traders - Newly listed markets before prices stabilize - Time-sensitive markets where news arrives faster than retail traders react

Markets where to avoid: - Major US presidential elections (everyone is watching, no edge) - Markets dominated by single resolution event (Super Bowl, World Cup finals) - Illiquid markets where you can't actually execute (look big, can't trade) - Markets you don't understand the resolution criteria for

## How to Practice +EV Trading

Step 1: Paper trade for at least one month. Our [paper trading guide](/blog/paper-trading-prediction-markets) covers how to make this useful. Track every trade with your reasoning. Did you actually find edge or were you guessing?

Step 2: Review your paper trade log. Where did you make money? Where did you lose? Were your probability estimates calibrated (when you said 70%, did you win 70% of those bets)?

Step 3: Start real trading with small positions. $5-25 per trade for the first month. Most "lessons" cost more than the trade size — make them cheap to learn.

Step 4: Track real performance. Spreadsheet, Predite tax report, whatever — but track. After 100 trades you'll have real signal. Before that, you don't.

Step 5: Adjust based on data, not feelings. If a category of trade is losing, stop trading it. If another is winning, scale up.

## A Concrete Example

You notice that a market on "Will Argentina win the 2026 Copa America?" is priced at 0.18 (18% probability). You're an Argentine soccer fan and have been watching their roster. You think the true probability is closer to 25%.

EV calculation: - Buy YES at $0.18 - If you're right and true prob is 25%: EV = (0.25 × $0.82) − (0.75 × $0.18) = $0.205 − $0.135 = +$0.07 per share - If you're wrong and true prob is 15%: EV = (0.15 × $0.82) − (0.85 × $0.18) = $0.123 − $0.153 = -$0.03 per share

The trade looks +EV based on your estimate. Sizing decision: - 7pp edge - Half-Kelly suggests 3-4% of bankroll - $1000 bankroll → $30-40 position - Buy 167-222 shares of YES at $0.18

If Argentina wins, you get $1.00 per share × 222 = $222 (profit ~$182). If they lose, you lose your $40. Played consistently over many similar trades, this kind of position sizing builds wealth.

## Common Mistakes by New +EV Traders

For a deeper dive, see our [10 mistakes guide](/blog/common-mistakes-new-prediction-traders).

- Treating one big win as proof the strategy works - Treating one big loss as proof it doesn't - Trading more often than your edge can support (over-trading kills returns) - Mistaking "I have a feeling" for "I have edge" - Ignoring base rates ("this is special, history doesn't apply") - Doubling down after losses to "recover" (path to bankruptcy)

## Bottom Line

Positive EV trading is the only sustainable approach to making money in prediction markets. It's not about being right on every trade — it's about being right on average and sizing positions to survive variance.

Master one market category. Track your probability estimates against actual outcomes (calibration matters). Size positions based on edge, not conviction. Track every trade for at least 100 trades before deciding if the strategy works for you.

The math is simple. The discipline is hard. Most people lose because they can't do the boring stuff consistently.