---
title: "Positive Expected Value in Prediction Markets: How to Find +EV Trades"
dek: "Learn how positive expected value works in prediction markets, why high-volume markets are harder to beat, and where mispriced contracts may appear."
description: "Learn how positive expected value works in prediction markets, why high-volume markets are harder to beat, and where mispriced contracts may appear."
author: "Matthew Figula"
author_role: "Prediction-market trader"
date: 2026-08-06
updated: 2026-08-07
category: "Guide"
tags: ["Edge", "Basics", "Execution", "Market structure"]
reading_time_minutes: 14
canonical: https://www.kosmos.fyi/blog/positive-expected-value-prediction-markets
---

# Positive Expected Value in Prediction Markets: How to Find +EV Trades

*Learn how positive expected value works in prediction markets, why high-volume markets are harder to beat, and where mispriced contracts may appear.*

Matthew Figula · Prediction-market trader · August 6, 2026 · Guide · 14 min read

**Positive expected value, or +EV, means that the price you are paying is lower than what you believe a contract is really worth.** If a YES contract costs 42¢ and you estimate that it has a 50% chance of paying $1, you have about 8¢ of expected value per contract before fees and trading costs.

That does not mean the trade is guaranteed to win. It means that if you could make the same type of trade many times with accurate probability estimates, you would expect to make money over time.

The formula is easy. The hard part is building a probability estimate that is better than the market's.

> **In one sentence**: A prediction-market trade is +EV when your estimate of the outcome's probability is higher than your all-in purchase price after accounting for the spread, fees, slippage, and uncertainty in your estimate.

## What is expected value?

Expected value is the average amount you would expect to make or lose if you repeated the same decision many times.

The standard formula is:

> **EV = (probability of winning × profit if correct) − (probability of losing × amount lost)**

Suppose a YES contract costs 52¢ and pays $1 if the event happens. You estimate that the event has a 60% chance of happening.

- If you are correct, your profit is 48¢: the $1 payout minus the 52¢ purchase price.
- If you are wrong, you lose the 52¢ you paid.

The calculation is:

> **EV = (0.60 × $0.48) − (0.40 × $0.52) = $0.08**

The trade has **8¢ of expected value per contract** before fees and other costs.

For a binary YES contract that you plan to hold until resolution, there is a simpler shortcut:

> **Expected value per contract = your estimated probability − your all-in price**

In this example:

> **60% − 52% = 8%**, or 8¢ per contract.

The same logic works for NO. Compare your estimated probability of NO with the all-in price of a NO contract.

![Expected value calculation for a 52-cent contract with a 60 percent estimated probability, producing 8 cents of expected value.](https://www.kosmos.fyi/blog/positive-expected-value-prediction-markets/expected-value-formula.webp)

*For a binary YES contract held to resolution, gross expected value is your probability estimate minus your all-in entry price.*

### A +EV trade can still lose

Suppose a YES contract costs 25¢ and you estimate that the real probability is 30%.

> **30% − 25% = +5%**

The trade is +EV, but it should still lose about seven times out of ten.

A good trade can lose. A bad trade can win. One result does not prove whether your original decision was smart. Expected value is useful because it evaluates the quality of the price, not just the outcome of one trade.

## The market price is only the starting point

Prediction-market prices are commonly read as implied probabilities. A YES contract trading at 63¢ suggests that the market is pricing the event at about 63%.

To have positive expected value, you need a reason to believe the contract is worth more than the price you can actually pay.

- If your fair probability is 70% and you can buy at 63¢, the gross edge is 7¢.
- If your fair probability is 60%, buying at 63¢ is negative EV—even if you think the event is more likely than not.

The important question is not simply:

> Will this happen?

It is:

> Will this happen more often than the current price implies?

### Use the price you can trade, not the headline price

The displayed market price may not be the price available to you.

For example, a market might show 51¢ because that is the midpoint between a 48¢ bid and a 54¢ ask. A buyer who wants an immediate fill must pay 54¢. A larger order might fill partly at 54¢, partly at 55¢, and partly at 57¢.

Your expected-value calculation should use:

> **Your average entry price at your intended size + fees + expected slippage**

If your probability estimate is 58%:

- At a displayed price of 51¢, the trade appears to have 7¢ of edge.
- At a 54¢ ask, the edge falls to 4¢.
- After another 1¢ of fees and slippage, the net edge is closer to 3¢.

A trade can look attractive on the market card and become average or negative once you inspect the order book.

## Why high-volume markets are usually harder to beat

High-volume prediction markets are usually more competitive. More people are watching them, more money can be traded, and pricing mistakes are more likely to be noticed.

That does not mean high volume makes a market correct. It means you normally need stronger evidence before claiming the market is wrong.

### 1. More traders are checking the same information

Major elections, Federal Reserve decisions, Bitcoin prices, and large sports markets attract constant attention. Traders are watching news, official data, related markets, and each other's orders.

When obvious new information arrives, several traders may reach the same conclusion at once. The price can adjust before a casual user even opens the market.

### 2. Professional traders can trade meaningful size

Sophisticated traders usually prefer markets where they can enter and exit without moving the price too much. High-volume markets are more likely to support larger orders, so they attract professional traders, market makers, and automated strategies.

These participants are not always right, but they increase the competition around obvious pricing errors.

### 3. News is usually priced faster

Active markets tend to process public information more quickly because more traders are paying attention and more capital is available to act.

A recent study of real-time prediction markets found that public signals were incorporated faster in liquid markets, while low-liquidity markets showed more underreaction and more price movement after the initial signal.

This creates a simple practical rule:

> The more active the market, the shorter the window for trading on obvious public news.

### 4. Related prices are compared constantly

A popular event may trade on Kalshi, Polymarket, sportsbooks, futures markets, or several related contracts. Traders can compare those prices and [trade any large difference](/blog/polymarket-arbitrage).

If one venue becomes clearly too cheap or too expensive, the gap may close quickly. In smaller markets, fewer people may be running those comparisons.

### 5. Better liquidity makes small edges tradable

High-volume markets often have tighter spreads and deeper order books. That is good for execution, but it also lets sophisticated traders act on small differences that would not be worth trading in a thin market.

Here is the basic trade-off:

| Market type | What is usually better | What is usually harder |
| --- | --- | --- |
| **High volume** | Tighter spreads, deeper books, easier entry and exit | More competition, faster reactions, smaller obvious errors |
| **Low volume** | Fewer eyes, slower updates, potentially larger errors | Wider spreads, less depth, harder exits, more uncertainty |

![Kosmos Market Tape snapshot comparing market odds, touch depth, 24-hour volume, and total volume.](https://www.kosmos.fyi/blog/positive-expected-value-prediction-markets/market-tape-volume-liquidity.webp)

*Volume measures past activity. Touch depth helps show how much can actually be traded near the current price.*

### Volume and liquidity are not the same thing

These terms are related, but they measure different things.

| Signal | What it tells you | What it does not tell you |
| --- | --- | --- |
| **Volume** | How much trading has already occurred | The price or size available right now |
| **Spread** | The gap between the best bid and ask | How much size exists behind those prices |
| **Depth** | How many contracts are available at different prices | Whether the current probability is fair |
| **Recent activity** | Whether attention and trading have increased | Whether the move came from informed trading |

A market can have high lifetime volume but a weak order book today. A newer market can have low total volume but reasonable liquidity at the price you want.

### High volume is not a truth score

Some volume comes from repeated turnover, uninformed trading, market-making activity, or platform incentives. Popular markets can still overreact, underreact, or reflect a biased crowd.

Research on prediction markets does not support the simple claim that more liquidity always produces a more accurate probability. An older study of TradeSports markets found cases where higher liquidity did not improve calibration and sometimes slowed the response to information.

The better conclusion is:

> **High volume usually means stronger competition and faster price discovery. It does not guarantee that the current price is correct.**

## Why low-volume markets can create more opportunity

Low-volume markets often receive less attention. There may be fewer informed traders, fewer market makers, and less automated monitoring. As a result, new information can take longer to affect the price.

This is where news, niche knowledge, and careful contract reading can matter most.

### Prices can stay stale after news

A smaller market may still reflect yesterday's information after a new poll, filing, injury report, court ruling, public statement, or data release changes the probability.

The presence of news is not enough. Three things must be true:

1. The news changes the probability of the exact contract.
2. The market has not already adjusted enough.
3. The remaining difference is larger than the spread, fees, slippage, and uncertainty in your estimate.

### One side can become inflated

Thin markets do not need much money to move. A small group of traders can push a contract several cents in one direction.

That move may be informed. It may also come from political bias, fan loyalty, social-media excitement, a misleading headline, or one trader who wanted an immediate fill.

This can create value on either side:

- The market may be **underreacting** to important information.
- The market may be **overreacting** to weak or emotional information.

Your job is to tell the difference.

### Niche knowledge matters more

Low-volume markets often reward traders who understand a narrow topic better than the general crowd.

Examples include:

- A local election with limited polling
- A bill with a complicated vote process
- A weather market tied to one exact station
- An entertainment market resolved by a specific chart or sales source
- A company event tied to an official filing
- A crypto market with an unusual deadline or threshold rule

In these markets, knowing the correct source and reading the resolution rules can matter more than building a complex model.

### But low volume adds real costs

A market can be badly priced and still be a bad trade.

Thin markets often have:

- Wide bid-ask spreads
- Little size available at the best price
- Large slippage on modest orders
- Difficulty exiting before resolution
- Greater resolution uncertainty
- A higher chance that the trader on the other side knows something you missed

Suppose you believe a contract is worth 50¢ and the displayed price is 40¢. That looks like a 10¢ edge. But the real ask is 45¢, only a small number of contracts are available there, and your estimate could reasonably be anywhere from 46% to 52%.

The apparent 10¢ opportunity may not be a real one.

## Sports betting has a clearer benchmark

In sports betting, many sportsbooks price the same game. That gives bettors a useful starting point for estimating fair value.

A common approach is to compare a recreational sportsbook with market-making books that accept sharper action, then remove the vig from the sharper prices. Pinnacle remains a common reference for major sports, but it is not the best source for every sport or market. Serious bettors often compare several market makers because sharpness varies across leagues, props, and bet types.

This makes sports +EV betting largely a comparison problem:

> **Sharp no-vig price → compare with the price offered elsewhere → calculate the edge**

Many non-sports prediction markets do not have that clean reference.

Kalshi and Polymarket may be the only direct prices for a unique political, economic, entertainment, weather, or geopolitical contract. Sometimes only one venue lists the exact event. Even when both platforms list something similar, their deadlines, wording, sources, or resolution rules may differ.

The prices also may not be independent. Many traders watch both venues and quickly trade large differences.

For unique prediction markets, you often need to build the baseline yourself.

| Market category | Useful inputs for estimating fair value |
| --- | --- |
| **Sports** | Market-making books, no-vig consensus, injuries, lineups, weather, closing movement |
| **Politics** | Polls, base rates, fundraising, endorsements, election rules, related markets |
| **Economics** | Futures, official release calendars, forecasts, prior data, central-bank statements |
| **Crypto and finance** | Spot price, volatility, options, time remaining, scheduled catalysts, nearby contracts |
| **Entertainment** | Eligibility rules, comparable releases, sales or chart data, source methodology |
| **Weather** | Official forecasts, model consensus, exact station, observation window, resolution source |

## How to find +EV trades in prediction markets

There is no single indicator that reveals the true probability of every event. A consistent research process is more useful.

### 1. Read the full contract

Start with the exact question, deadline, source, threshold, and resolution process.

“Will Bitcoin touch $100,000 before December 31?” is different from “Will Bitcoin close above $100,000 on December 31?” One requires a single intraday touch. The other depends on the price at one specific time.

A correct view of the general topic can still produce a losing trade if you misunderstand what the contract measures.

### 2. Check volume, spread, and depth

Do not judge a market from volume alone. Check:

- Total and recent volume
- Best bid and best ask
- Spread
- Size available at the best price
- Depth behind that price
- Recent movement and reversals

This tells you whether the market is active and whether the quoted opportunity is actually tradable.

### 3. Build an independent probability range

Use external evidence to estimate a reasonable range before anchoring too heavily on the market price.

For example:

> **Estimated fair range: 42% to 48%**

A range is more honest than claiming the event is exactly 45.3% likely. If the market sits inside your range, you probably do not have a strong edge.

### 4. Follow news and upcoming catalysts

List the events that could change the probability before the market closes: data releases, votes, court decisions, polls, earnings, speeches, injuries, weather updates, or official announcements.

Then ask:

- Does this information affect the exact contract?
- How much should it change the probability?
- Has the market already moved?
- Is the source credible and new?

A relevant headline is not automatically an edge. It becomes useful only when the price has not adjusted enough.

![Four-step checklist for evaluating whether a news headline creates value in a prediction market.](https://www.kosmos.fyi/blog/positive-expected-value-prediction-markets/news-to-market-impact.webp)

*A headline matters only when it changes the exact contract, comes from reliable evidence, has not been fully priced in, and leaves enough edge after costs.*

### 5. Compare venues and related markets

Compare Kalshi and Polymarket when both list the event, but first confirm that the contracts really match.

Then look for other markets that measure part of the same outcome. A Fed contract may connect to interest-rate futures. A Bitcoin threshold market may connect to spot price, volatility, options, and nearby strike markets. An election contract may connect to polling and nomination markets.

Related markets help you test whether your probability estimate is internally consistent.

### 6. Separate direction from value

A directional view tells you whether the underlying situation is improving or worsening. It does not tell you whether the contract is cheap.

You can be bullish on Bitcoin and still think a Bitcoin YES contract is overpriced. You can think a candidate is gaining momentum and still believe the market has already moved too far.

> **Direction asks where the probability is moving. Expected value asks whether the current price is still wrong.**

Directional signals, asset moves, and technical indicators can support a probability estimate. They cannot replace one.

![Illustrative Bitcoin trend compared with a prediction-market contract to show that direction is not the same as expected value.](https://www.kosmos.fyi/blog/positive-expected-value-prediction-markets/directional-analysis-not-ev.webp)

*A bullish asset view can still produce a negative-value trade when the related contract is already too expensive.*

### 7. Calculate net EV and require a margin of safety

For a YES contract held to resolution:

> **Net EV ≈ your probability − average entry price − expected fees and slippage**

Suppose your fair probability is 45%:

- Displayed price: 36¢
- Best ask: 39¢
- Expected fees and slippage: 1¢

Your rough net EV is:

> **45¢ − 39¢ − 1¢ = +5¢ per contract**

Your probability estimate is also uncertain. Treat that uncertainty as a cost.

If your fair range is 42% to 48% and the all-in purchase price is 41%, the trade may be slightly positive EV, but the evidence is weak. At an all-in price of 34%, the margin is much stronger.

The smaller and more unusual the market, the larger your margin of safety should usually be.

## A simple low-volume example

Consider a hypothetical market:

> **Will a bill pass the Senate by June 30?**

The market shows YES at 34¢ and has relatively low volume. When you open the book, the best ask is actually 38¢ and only modest size is available.

You review:

- The current public vote count
- Statements from undecided senators
- The legislative calendar
- The exact meaning of “pass” in the contract
- Whether the market requires only Senate passage or final enactment
- Related contracts on the same bill

New reporting suggests that two previously undecided senators are likely to support the bill. The prediction-market price has barely moved.

You estimate a fair range of 42% to 48%, with a midpoint of 45%.

At a 38¢ ask, the gross expected value is about 7¢:

> **45¢ − 38¢ = +7¢**

After allowing 1¢ for fees and slippage, the rough net EV is 6¢ per contract.

That may be a real opportunity—but only if the reporting is credible, the contract rules match your interpretation, and you can buy near 38¢ without pushing the price much higher.

This is the basic low-volume opportunity: meaningful information arrives, but the market updates slowly.

It is also the basic low-volume risk: one wrong assumption about the source, rules, or available price can remove the entire edge.

## Common mistakes when looking for +EV

### Treating every low-volume market as soft

Some small markets are mispriced. Others are simply expensive to trade, difficult to analyze, or dominated by one better-informed participant.

### Assuming high volume means the market must be right

High volume usually means more competition. It does not remove bias, noise, or overreaction.

### Using the displayed price instead of the ask

Use the average price you can actually receive at your intended size.

### Treating relevant news as unpriced news

A headline can be important and still provide no value if other traders have already reacted.

### Ignoring resolution rules

The exact wording determines what pays. Your broader opinion about the topic does not.

### Using false precision

A realistic probability range is usually more useful than a highly precise estimate based on weak inputs.

### Judging the process from one result

A +EV trade can lose. Measure the quality of your prices and forecasts over many decisions.

## How Kosmos helps with the research process

Kosmos brings the main parts of prediction-market research into one workspace:

- **Markets** compares contracts across Kalshi and Polymarket and shows venue-level price and volume.
- **News & market impact** connects current events with the markets they may affect.
- **Market Tape** includes movement, reversal, spread, touch depth, recent volume, total volume, trend strength, and realized variation.
- **Assets** adds live data for stocks, crypto, ETFs, foreign exchange, and commodities.
- **Traders** shows track-record metrics including P&L, ROI, win rate, volume, and calibration.
- **Agents** can research markets, explain moves and catalysts, and inspect contract wording with supporting evidence.

These tools do not guarantee positive expected value. They help you check the inputs that determine whether an apparent edge is real.

A practical Kosmos workflow is:

> **Find a market → read the contract → check volume and execution → review news and related data → build an independent probability → calculate net EV**

![Five-step workflow for researching positive expected value in a prediction market.](https://www.kosmos.fyi/blog/positive-expected-value-prediction-markets/kosmos-positive-ev-workflow.webp)

*A repeatable process starts with the contract, builds an independent probability range, compares venues, checks execution, and records the thesis.*

## Final takeaway

Positive expected value is the gap between what a contract costs and what it is worth based on your best probability estimate.

High-volume markets usually have better execution and faster price discovery, but they also attract more sophisticated competition. Obvious mistakes tend to disappear quickly.

Low-volume markets can contain larger pricing errors because fewer people are watching, a small group can push one side too far, and niche information may be missed. They also come with wider spreads, lower depth, harder exits, and more uncertainty.

The goal is not to trade every small market or avoid every large one. It is to find situations where your information and analysis are stronger than the current price—and where the edge remains after execution costs and uncertainty.

## Sources and methodology

- [Google Search Central: Creating helpful, reliable, people-first content](https://developers.google.com/search/docs/fundamentals/creating-helpful-content)
- [Smarkets: How to calculate expected value in betting](https://help.smarkets.com/hc/en-gb/articles/214554985-How-to-calculate-expected-value-in-betting)
- [Polymarket documentation: Prices and order book](https://docs.polymarket.com/concepts/prices-orderbook)
- [Kalshi Help Center: The order book](https://help.kalshi.com/en/articles/13823828-the-orderbook)
- [Angelini and De Angelis: When Do Markets Fully Process Public Information?](https://arxiv.org/abs/2606.07811)
- [Paul C. Tetlock: Liquidity and Prediction Market Efficiency](https://business.columbia.edu/sites/default/files-efs/pubfiles/3098/Tetlock_SSRN_Liquidity_and_Efficiency.pdf)
- [Kalshi Help Center: Market maker program](https://help.kalshi.com/en/articles/13823819-how-to-become-a-market-maker-on-kalshi)
- [Polymarket documentation: Market making](https://docs.polymarket.com/trading/market-making)
- [Unabated: Who sets the sports-betting line?](https://unabated.com/post/who-sets-the-sports-betting-line-market-makers)
- [Kosmos Markets](https://app.kosmos.fyi/markets)
- [Kosmos News](https://app.kosmos.fyi/news)
- [Kosmos Market Tape](https://app.kosmos.fyi/analytics)
- [Kosmos Traders](https://app.kosmos.fyi/traders)
- [Kosmos Agents](https://app.kosmos.fyi/agents)

> **Disclaimer**: This article is for educational purposes only and is not financial advice. Prediction-market trading involves risk, including the risk of losing the full amount paid for a contract.

## Frequently asked questions

### What does +EV mean in prediction markets?

+EV means positive expected value. A trade is +EV when your estimated probability of receiving the winning payout is higher than the break-even probability implied by your all-in purchase price.

### Are high-volume prediction markets more accurate?

They are often harder to beat because more traders, market makers, and automated systems are competing to correct errors. However, volume alone does not guarantee accuracy. Popular markets can still be biased or slow to process some information.

### Are low-volume prediction markets easier to trade profitably?

They can contain larger pricing errors because fewer people are watching them, but they are not automatically easier. Wide spreads, low depth, slippage, resolution risk, and informed counterparties can erase the apparent edge.

### How do I estimate the fair probability of a prediction-market event?

Start with the contract rules, base rates, official data, related markets, current news, and known catalysts. Build a probability range, then compare it with the executable market price rather than the displayed midpoint.

### Can a positive-EV trade still lose?

Yes. Expected value describes the average result across many similar decisions. Any individual +EV trade can lose.

### How do fees and liquidity affect expected value?

Fees, spread, and slippage increase your real entry cost. Use the average price you expect to receive at your intended size and subtract all expected trading costs before deciding whether the trade remains +EV.