The world’s first prediction market for private, unlisted companies.

NOTSTOCKS
Sign in
Back to blog

Beginners: How Prediction Markets Work and Trade Private Markets

Prediction market contract title card

Prediction markets are exchanges where traders buy and sell contracts tied to the outcome of a future event, and the contract’s price doubles as the crowd’s estimate of how likely that event is. A contract trading at $0.70 implies roughly a 70% chance the event happens. When the event resolves, winning contracts settle at $1 and losing ones at $0, which is why these markets double as live, aggregated forecasts.

TL;DR:

  • High liquidity, clear resolution rules, and active surveillance greatly enhance confidence in prediction market prices.
  • Small trade sizes and thorough tracking of own calibration improve trading results and understanding of market microstructure.
  • Biases such as herding and favorite-longshot effects can distort prices, especially in thin or highly volatile markets.
  • Regulation by authorities like the CFTC ensures transparency, dispute procedures, and protections but does not guarantee profits.
  • Private company prediction markets, like NotStocks, leverage AI signals and visible trade histories to make opaque markets more accessible.
Notstocks
Explore Private Company Prediction Markets
NotStocks lets you trade predictions on private companies across healthtech, fintech, AI, and other emerging sectors.
Explore NotStocks

Table of Contents

What prediction markets are and key concepts

A prediction market lets people trade contracts on a specific, defined event: will a bill pass, will a company hit a revenue target, will a storm make landfall by a certain date. Each contract represents one possible outcome. In a binary market there are exactly two: “yes” and “no.” In a multiple-choice market, several outcomes compete, and only one ultimately pays out.

Prediction contract outcomes and payouts

Ownership of a contract is simple to track. You buy a “yes” share at the current market price, and your potential profit is the gap between what you paid and what the contract settles at. If the event resolves in your favor, the CFTC’s overview of event contracts confirms that regulated contracts settle at $1 for the correct outcome and $0 for the incorrect one.

A simple example makes this concrete: say you buy a “yes” contract at $0.60. If the event happens and the market settles at $1, you collect $0.40 in profit per contract. If it does not happen, the contract settles at $0 and you lose your $0.60 stake.

A few terms come up constantly:

  • Event contract: a tradable security whose value depends entirely on a future, verifiable outcome.
  • Settlement: the final payout once the event resolves, typically $1 or $0 for binary contracts.
  • Implied probability: the market price read as a percentage chance, so a $0.60 price implies a 60% chance.

This structure is what separates a prediction market from a simple bet: the price moves continuously as new information arrives, not just at the moment of settlement.

How pricing and payouts work

The price of a contract is the fastest way to read a market’s current view of probability. A $0.35 “yes” price implies a 35% chance of that outcome, and that number updates every time someone trades. Research from Stanford on prediction markets in theory and practice found that market-generated forecasts often outperform many conventional benchmarks precisely because the price mechanism pools scattered information from many traders into one number.

Prediction markets have been shown to aggregate dispersed information into forecasts that often beat standard benchmarks. That’s the core argument for why prices are worth watching, not just trading.

Turning a price into an expected profit takes one more step. If you buy a contract at $0.40 and it settles at $1, your gross profit is $0.60 per contract, before fees. If it settles at $0, you lose your full $0.40 stake. The math is straightforward, but two frictions eat into it in practice:

  • Bid-ask spreads: the gap between what buyers are willing to pay and what sellers will accept, which widens in thin markets and reduces your effective entry or exit price.
  • Trading fees: a small percentage charged on activity that lowers your realized return even when your forecast is right.

Not every contract is a clean binary. Some markets use partial-payout or scalar contracts, where the settlement value lands somewhere between $0 and $1 depending on where the actual outcome falls within a range, such as an economic indicator landing between two thresholds. These contracts behave more like a weighted average than a coin flip, and they reward traders who can estimate a range accurately rather than just a direction.

Market mechanics: AMM vs CLOB, liquidity, and price impact

Two different trading infrastructures sit behind most prediction markets, and the choice between them changes how closely your profit tracks your accuracy.

An automated market maker, or AMM, prices contracts using a cost function rather than matching individual buyers and sellers. Early AMM designs tied a trader’s profit directly to moving the price toward their true belief: if you were right and you moved the price in that direction, you were mathematically guaranteed a profit under the right conditions. A liquidity-sensitive AMM design from Carnegie Mellon researchers builds on this by adjusting how much the price moves per trade as volume grows, making deeper markets less elastic and allowing the market maker to manage worst-case losses.

AMM liquidity and price impact paths

A central limit order book, or CLOB, works differently: it matches resting buy and sell orders directly, the way a stock exchange does. Here, the link between being right and making money is less automatic. An analysis of CLOB-based prediction markets found that bid-ask spreads, limited liquidity, and price impact can cause even accurate forecasters to lose money, because execution quality matters as much as the underlying call.

A few mechanics shape what you actually experience as a trader:

  • Liquidity determines how much size you can trade without moving the price against yourself.
  • Slippage is the difference between the price you expected and the price you got, which grows with trade size in thin markets.
  • Spread is the built-in cost of trading immediately rather than waiting for a better price.

For forecasters, the design choice matters: AMMs tend to reward being early and directionally correct, while CLOBs reward being correct and able to execute efficiently.

Pro Tip: Start with small trade sizes in any new market until you’ve watched how much the price moves per dollar traded.

Types of contracts and example markets

Recognizing the contract format tells you what kind of forecast you’re actually making.

  1. Binary contracts settle at $1 for “yes” or $0 for “no,” the simplest and most common format, used for elections, sports, and yes/no corporate events.
  2. Multiple-choice contracts split one event into several competing outcomes, such as “which candidate wins,” where only one option ultimately pays out.
  3. Scalar or range contracts settle somewhere between $0 and $1 depending on where the real-world number lands within a defined range, common for economic data releases.
  4. Partial-payout contracts blend elements of the above, useful when an event has graduated outcomes rather than a clean win or loss.

Whatever the format, the contract is only as good as its resolution language. A market asking “will inflation exceed 3%” needs to specify which index, which month, and which data source settles it. The CFTC’s guidance on event contracts emphasizes that regulated markets require clear, objective resolution criteria precisely because ambiguity is the leading cause of disputes.

Common real-world categories include elections and political outcomes, scheduled economic releases like employment or inflation reports, corporate milestones such as product launches or earnings thresholds, and weather events tied to an official forecast source.

How to trade: a step-by-step checklist for beginners

Getting started is less about finding an edge on day one and more about understanding the mechanics before you put money behind an opinion.

  1. Pick a reputable, clearly regulated market and read its rulebook before trading a single contract.
  2. Check liquidity and recent volume for the specific contract, not just the platform overall, since thin markets can swing on a single trade.
  3. Use limit orders rather than market orders when possible, so you control the price you pay instead of accepting whatever is available.
  4. Size your stakes deliberately, treating each trade as one bet among many rather than a single high-conviction swing.
  5. Diversify across uncorrelated events so one wrong call doesn’t define your results.
  6. Keep a simple log of your trades, your stated probability at entry, and the actual outcome, so you can check your own calibration over time.

If you were right only four times, your confidence is running ahead of your accuracy.*

The habit of recording and reviewing trades matters more than any single pick. It’s the fastest way to learn whether you’re reading prices well or just reacting to headlines.

Risks, biases, and common pitfalls

Prediction markets are informative, not infallible, and several failure modes recur often enough to plan around.

Manipulation and insider information are the most direct threats: a trader with private knowledge of an outcome, or one willing to push a large order through a thin market, can distort the price temporarily. Ambiguous resolution language compounds this, since a disputed outcome can leave a market stuck or settled in a way participants didn’t expect.

  • Low liquidity produces noisy prices that swing on small trades and don’t reflect genuine consensus.
  • Favorite-longshot bias can distort prices at the extremes, where longshot outcomes sometimes trade above their true probability and favorites slightly below.
  • Herding causes traders to pile into a price move rather than independently evaluating the event, amplifying swings that have little to do with new information.

Calibration also varies by domain: research on prediction market biases has found that political markets often show underconfidence and a tendency to compress toward 50% over time, while economic or firm-specific markets behave differently depending on how much public data traders have to work with. This is detailed in research on calibration and domain-dependent biases in prediction markets, and it’s a useful reminder that a price near 50% isn’t always a coin flip. It can also be a market still absorbing information, or one where traders are systematically underconfident.

Treat any single price as a snapshot of current sentiment, not a guarantee, and weight it more heavily when volume and liquidity are high.

Regulation and settlement: what to check

Before trading any real money, it’s worth knowing what regulatory oversight actually buys you.

In the United States, the CFTC’s framework for event contracts requires exchanges to meet market integrity standards, including surveillance for manipulation and clear procedures for how disputes get resolved. Regulated markets have existed for decades under this kind of oversight, used both to hedge and to speculate on event outcomes.

A few concrete things to look for before you trade on any platform:

  • A published rulebook that defines exactly how and when a contract resolves, down to the data source used.
  • Evidence of market surveillance, meaning the exchange actively monitors for manipulation rather than relying purely on volume to self-correct.
  • A clear dispute procedure for what happens when an outcome is contested or a data source is unclear at settlement time.
  • Customer protections around how funds are held and how settlement payouts are processed.

None of this guarantees a profitable trade. It does mean that when a market resolves in a way you disagree with, there’s a documented process rather than a shrug.

Real-world applications: forecasting, hedging, and corporate use cases

Prediction markets show up well beyond casual speculation, embodying the principle to navega con datos, no con ruido.

  • Election and economic forecasting: researchers and journalists track market-implied probabilities as a running estimate of how a race or a policy decision is likely to unfold.
  • Corporate internal markets: some companies run internal prediction markets to forecast product launch timing or sales targets, pooling employee knowledge that might otherwise stay siloed.
  • Hedging event risk: businesses exposed to a specific weather pattern or regulatory decision can use event contracts to offset some of that exposure.
  • Academic and policy research: economists cite market-generated probabilities as a data point alongside polling and expert judgment, since the Stanford research on prediction markets found these forecasts frequently hold up against other methods.

Compared to a poll, a prediction market updates continuously and carries a real cost to being wrong, which gives participants a direct incentive to trade on genuine information rather than stated preference. Compared to an AI model’s forecast, a market reflects what humans are actually willing to put money behind right now, which can pick up information a model trained on historical data hasn’t seen yet.

How NotStocks applies prediction-market mechanics to private companies

Most of what’s described above was built for public, well-documented events: elections, economic releases, sports. NotStocks applies the same mechanics to a category that regulated prediction markets rarely touch: private, unlisted companies.

The platform lets traders speculate on the future success of startups and private enterprises across sectors like healthtech, fintech, and AI, using event contracts the same way you’d trade a binary or scalar market elsewhere. Because private companies don’t publish the earnings reports or regulatory filings that public-market forecasters rely on, NotStocks adds AI-based signal analysis for each listed company and keeps a visible price and trade history, aiming to make an otherwise opaque market somewhat easier to read.

  • The platform reports significant trading volume across its markets within a 24-hour period, with multiple companies listed.
  • Anyone can submit or claim a company listing, and claimed listings share in that market’s trading fees.
  • A 1% trading fee applies to activity on the platform.

You can see this in practice on an individual company listing page, where price history and valuation context sit alongside the current market price, or in the platform’s broader guide to prediction markets for private companies.

When to trust a market, and when to stay skeptical

A price is only as good as the liquidity behind it. Thin volume, vague resolution rules, or no regulatory oversight should all lower your confidence in a number, no matter how official it looks on screen. Thick volume, a clear rulebook, and active surveillance should raise it.

The honest starting posture is conservative: trade small, watch how prices move relative to your own estimate, and treat the first few markets as tuition for learning the microstructure, not as a shortcut to easy returns.

— Max

Try prediction markets on private companies with NotStocks

Everything above applies to public events with public data. Private companies are the harder case: less disclosure, fewer comparable filings, and a smaller pool of traders setting the price. NotStocks was built specifically for that gap, giving traders a place to speculate on the future of startups and private enterprises across healthtech, fintech, and AI, backed by AI-generated signal analysis and a visible price history for each listing.

Notstocks

If you run or know a private company, you can claim its listing for free and share in the trading fees generated on that market. If you’re here to trade, start with the main NotStocks platform, where a 1% fee applies to trading activity, or browse the learning hub for a deeper walkthrough of how private-company markets differ from the public ones covered above. For a sense of what a live market looks like, the guide to private-company prediction markets is a reasonable next stop.

This article is general information, not a substitute for advice from a qualified financial advisor. Consult a qualified financial professional about your own circumstances before acting on anything here.

FAQ

Can you actually make money on prediction markets?

It’s possible to profit if your probability estimates are consistently better calibrated than the market’s, but fees, spreads, and liquidity limits all cut into realized returns. Research on order-book prediction markets found that even accurate forecasters can lose money once spread and price impact are factored in, so accuracy alone doesn’t guarantee profit.

How do prediction markets pay out?

Binary contracts settle at $1 for the correct outcome and $0 for the incorrect one, as described in the CFTC’s event contract overview. Scalar or partial-payout contracts settle somewhere between $0 and $1 depending on where the actual result falls within a defined range.

What are the top prediction market platforms?

Platforms vary by focus: some concentrate on politics and public events, others on sports, and some, like NotStocks, focus specifically on private, unlisted companies. The right one depends on what you want to forecast and whether you need regulated oversight or a niche category like private-market speculation.

How do you improve your results in prediction markets?

Start by tracking your own calibration: record the probability you assigned at entry and compare it to actual outcomes over time. Favor markets with clear resolution rules and enough liquidity that your trade size doesn’t move the price against you.

Are prediction markets regulated?

In the United States, the CFTC oversees certain event contracts and requires exchanges to meet market integrity and surveillance standards. Oversight varies by country and by platform, so it’s worth checking a specific market’s regulatory status before trading.

Sources