5 Prediction Market Concepts That Confuse Beginners

Understanding prediction market concepts for beginners starts with recognizing these five specific misunderstandings. In November 2024, a market on Polymarket asked: "Will Trump be president-elect by midnight?" At 11:47 PM EST, it traded at 94¢. A beginner deposited $500 and bought 531 shares at 94¢ each. Thirteen minutes later, the Associated Press called the race. The market resolved "Yes." Those shares paid $531. Profit: $31, or 6.2%.

The beginner posted on Reddit: "Only made $31 on a 94% sure thing. What's the point?"

A 94% market isn't a sure thing. It's a 94-cent dollar with a 6% chance of worthlessness. That beginner confused price with certainty, mistook high probability for guaranteed profit, and didn't realize that a 6.2% return in thirteen minutes would annualize to something absurd if repeatable (which it isn't, but that's a separate lesson).

Prediction markets use familiar concepts (percentages, buying and selling, odds) but combine them in deceptively simple ways. Five specific concepts trip up most newcomers. Not because they're mathematically complex, but because they look like things you already understand from sports betting, stock trading, or polling data.

Understanding these fundamentals is one of the core topics we unpack across the learn hub, where we break down prediction market terminology in plain English. By the end of this piece, you'll know exactly what each concept means, why it matters when you're deciding whether to place a trade, and which mistake to avoid.

---

1. Understanding Prediction Market Odds: Probability vs Odds vs Price

A 62¢ price isn't the same as a 62% poll or 62% odds. Here's the difference that matters when you trade.

Probability is the market's belief expressed as a percentage. When you see 62%, the market is saying there's a 62-in-100 chance of this outcome.

Odds are the ratio of outcomes. Understanding prediction market odds requires separating probability (market belief), odds (return ratio), and price (entry cost). A 62% probability converts to 1.61-to-1 decimal odds, or +161 in American odds (the format sportsbooks use). Odds show potential return relative to stake.

Price is what you pay per share. On prediction markets, a "Yes" share trading at 62¢ means you pay $0.62 for a contract that pays $1.00 if the outcome happens. The price is the implied probability, expressed in cents-per-dollar.

If you buy 100 shares at 62¢, you spend $62. If "Yes" resolves, you get $100. Profit: $38. If "No" resolves, you get $0. Loss: $62.

This distinction matters for understanding prediction market odds: the same 62% probability converts to +161 American odds or 1.61 decimal odds, depending on how you want to calculate potential returns.

Why this matters when you're deciding to trade

Probability tells you the market's belief. Odds tell you potential return. Price tells you entry cost. And critically, the displayed percentage is the market's current aggregate belief, which may be wrong.

The market price reflects what traders collectively think, weighted by how much capital they're willing to risk. A 62% market doesn't mean experts have calculated a 62% chance. It means the marginal trader (the last person to buy or sell) thought 62¢ was fair value for a dollar that pays if this outcome occurs.

That price can be wrong for several reasons:

• Incomplete information: Traders don't know something you know
• Emotional bias: People overreact to recent news or bet on favorites
• Strategic positioning: Large traders moving markets for reasons unrelated to probability
• Illiquidity: So few trades that the price reflects only a handful of opinions

Real Example: Bitcoin at $100K

In December 2024, a Polymarket question asked: "Will Bitcoin hit $100,000 before January 1, 2025?" On December 5th, Bitcoin traded at $96,800. The market sat at 73¢.

• Probability: 73% (meaning roughly three-in-four parallel universes see Bitcoin cross $100K)
• Decimal odds: 1.37 (meaning you win $1.37 for every $1 wagered)
• American odds: -270 (meaning you'd risk $270 to win $100)
• What you pay: 73¢ per share, winning $1.00 if correct

A beginner saw "73%" and thought: Bitcoin's already at $96K. It only needs to go up 3.3%. This is free money.

Bitcoin peaked at $99,830 on December 17th. It never crossed $100,000 in 2024. The market resolved "No." That beginner lost their entire stake.

The lesson: A high percentage doesn't mean "likely to happen soon" or "safe bet." It means the market has priced in a 73% chance. Your edge exists only if you believe the true probability is higher than 73%, and you have reasons the market doesn't.

For more on how these percentages translate into actionable bets, see what the price actually means.

---

2. Liquidity: How Market Depth Affects Your Actual Trading Costs

Liquidity isn't total volume. It's how much you can trade right now without moving the price.

Beginners see "$2.3 million volume" on a market and assume they can instantly buy or sell any amount at the displayed price. That's not how prediction markets work. Volume shows historical activity. Liquidity shows current tradability.

Imagine selling a rare collectible. Someone once paid $10,000 for a similar item (that's "volume"). But if you try to sell yours today and there's only one buyer offering $7,000, you're facing a liquidity problem. The market exists at a price you don't want.

The order book determines what you actually pay

Prediction markets operate on an order book, a live list of bids (buy orders) and asks (sell orders) at different price levels. When you see "Trump 62%," that's the midpoint between the best bid and best ask. It's not necessarily the price you'll get.

Here's a simplified order book for "Trump wins 2024":

Sell orders (asks):

• 62.1¢ — 500 shares available
• 62.2¢ — 1,200 shares available
• 62.5¢ — 3,000 shares available

Buy orders (bids):

• 61.9¢ — 800 shares available
• 61.7¢ — 2,500 shares available

If you want to buy 600 shares right now using a market order (meaning you'll accept whatever price is available):

• First 500 shares at 62.1¢ = $310.50
• Next 100 shares at 62.2¢ = $62.20
• Total cost: $372.70
• Average price paid: 62.12¢

You paid 0.12¢ more per share than the displayed "62¢" because you exhausted the best price level and had to tap into the next tier.

Now imagine trying to buy 5,000 shares. You'd blow through 62.1¢, 62.2¢, and 62.5¢, paying an average much higher than 62¢. This is called slippage, the gap between the price you expected and the price you actually paid. It happens when there isn't enough liquidity at your target price to fill your entire order.

High liquidity vs low liquidity in action

Compare two markets from February 2025:

Polymarket: "2024 Presidential Election Winner" (after election but before inauguration)

• Bid/ask spread: 99.8¢ / 99.9¢ (a 0.1¢ gap, meaning almost no friction)
• Depth: Over $100,000 available within 0.2¢ of midpoint
• Slippage on $10,000 order: ~0.05¢ per share (meaning you lose almost nothing to execution cost)

Polymarket: "Will Senator X resign by March 15?"

• Bid/ask spread: 12¢ / 18¢ (a 6¢ gap, 50% of the price)
• Depth: $340 available at best bid, $180 at best ask
• Slippage on $10,000 order: Impossible to fill without moving price 10+ cents

In the first market, you can trade size without penalty. In the second, the moment you try to buy $1,000 worth of shares, you'll push the price from 18¢ to 25¢ or higher. When you try to sell later, the bid might be 8¢. You've locked in a loss just from trading costs.

How low liquidity destroys beginner exits

When Kalshi launched congressional election markets in October 2024, early volume was thin. A trader bought $800 of "Yes" shares in a House race at 34¢. Two weeks later, polling shifted and the true probability was closer to 50%. That trader tried to sell. The bid was 29¢. Despite being "right" about the direction, they couldn't exit profitably because liquidity had dried up.

Common misconception: "I can always sell at the current price."

Reality: You can sell at the current bid, which may be significantly lower than the last trade price or the displayed midpoint. In illiquid markets, that gap eats 10-20% of your position value.

Practical tip: Before trading, click into the order book (most platforms show this as "Depth" or "Order Book"). If the spread is wider than 2-3¢, or if there's less than $5,000 in depth on each side, you're paying a tax to enter and exit. That tax destroys edge unless you're extremely confident in your forecast.

To understand more about how these mechanics fit into our complete guide to prediction markets, including platform-specific features, see the broader ecosystem overview.

---

3. Market Resolution: Who Decides What Happened, and What If They're Wrong?

In March 2024, a Polymarket market asked: "Will Zelenskyy wear a suit during his next public appearance?" It had $14 million in volume. Zelenskyy appeared at a press conference wearing a dark blazer, collared shirt, no tie. Some users argued it was a suit jacket. Others said it was a sport coat. The market rules stated: "Resolves YES if Zelenskyy wears a traditional business suit (matching jacket and trousers)."

The trousers were olive green. The jacket was navy. Those who bet "Yes" claimed the definition was ambiguous. Those who bet "No" said the rules clearly required matching pieces. The dispute went to UMA Protocol, a decentralized oracle where token holders vote on resolution. After three days of voting, UMA resolved it "No." The "Yes" holders lost $6.2 million.

This wasn't market manipulation or platform error. It was resolution ambiguity, when the outcome seems obvious to you but the exact wording of the rules allows for multiple interpretations.

How resolution actually works

Resolution is the official determination of a market's outcome, which triggers payouts. Every market has predefined resolution criteria written into the contract. The quality of those criteria determines whether the market resolves cleanly or turns into a dispute nightmare.

The real question beginners need answered: Can I get screwed by a bad resolution decision?

Yes, if the rules are vague. Here's how each platform handles it:

Polymarket uses UMA Protocol, where token holders vote if someone challenges the initial resolution. If you think the resolution is wrong, you post a $10,000 bond in UMA tokens to force a vote. If the community sides with you, you get the bond back and the resolution flips. If they side against you, you lose the bond. This system discourages frivolous disputes but doesn't eliminate subjective calls when the rules are poorly written.

Kalshi operates as a CFTC-regulated exchange. Markets resolve based on official sources specified upfront: Federal Reserve data, Bureau of Labor Statistics reports, Treasury Department announcements. There's no community vote. Kalshi's operations team applies the stated rule to the stated source. If you dispute, it goes to regulatory arbitration, not a token vote.

PredictIt is run as an academic project by Victoria University of Wellington. University researchers make resolution calls based on stated rules. Disputes go to the university administrators. There's no formal appeals process beyond contacting the research team.

When resolution goes right

In summer 2023, Polymarket offered: "Will Donald Trump be indicted in Georgia by December 31, 2023?"

Resolution criteria: "Resolves YES if an indictment document is filed in Fulton County Superior Court listing Donald J. Trump as a defendant before 11:59 PM EST on December 31, 2023. Source: Official Fulton County court records."

On August 14, 2023, at 9:37 PM, a grand jury indicted Trump. The court clerk's office posted the document online at 11:04 PM. The market resolved "Yes" within 22 minutes of the filing timestamp appearing in the public record.

No dispute. No ambiguity. The resolution criteria specified an objective source and a clear threshold. Traders knew exactly what event would trigger resolution.

What can go wrong with vague resolution criteria

Vague criteria: "Will X be considered a success?" (Success according to whom?)

Ambiguous timing: "By end of year" (In what timezone? Does midnight count?)

Source conflicts: "Per major news outlet" (What if outlets disagree?)

Subjective judgment: "Will X's speech be well-received?" (Well-received by whom? Measured how?)

In December 2024, a Polymarket market asked: "Will Joe Biden use the word 'democracy' more than 10 times in his State of the Union?" Biden gave the speech. Transcripts varied. The official White House transcript listed 11 uses. C-SPAN's transcript listed 9. The market rules didn't specify which transcript was authoritative. UMA voters sided with the White House version. "Yes" won. But traders who relied on C-SPAN lost money, not because their count was wrong, but because the resolution source wasn't clear.

Practical tip: Read the resolution criteria before you trade. If it includes subjective terms like "widely regarded," "generally considered," or "major news outlets," skip the market. You're betting not just on the outcome, but on how the outcome will be measured. If the measurement is ambiguous, you're exposed to resolution risk that has nothing to do with your forecast.

For a deeper dive into how contracts get resolved across platforms, see how contracts get resolved.

---

4. Expected Value vs Actual Outcome: Why Good Bets Lose All the Time

In 2016, FiveThirtyEight's model gave Donald Trump a 28.6% chance of winning the presidency. He won. Does that mean the model was wrong?

No. It means an event with roughly 3-in-10 odds occurred. If you ran that election 100 times in parallel universes with identical conditions, Trump should win about 29 times. This one happened to be one of the 29.

A single outcome doesn't validate or invalidate a probability. A 70% bet that loses wasn't necessarily a bad bet. You just hit the 30%. A 15% longshot that wins wasn't necessarily a good bet. It might've been a terrible bet that got lucky.

How to think about expected value before placing a trade

Start with a concrete scenario. You buy "Yes" shares at 62¢. You believe the true probability is 70% (higher than the market's implied 62%). Here's what that means in practice:

If you make this exact trade 100 times:

• You'll win roughly 70 times, profiting 38¢ per win = $26.60 in profits
• You'll lose roughly 30 times, losing 62¢ per loss = $18.60 in losses
• Net expected result: +$8.00 across 100 bets

That +$8 per 100 bets is your expected value (EV), the average outcome if you repeated this decision many times. Understanding prediction market odds through expected value calculations reveals why 70% bets at 62¢ beat 90% bets at 88¢. The formula distills to:

EV = (Your estimated probability × Payout) - Cost

In this case: (0.70 × $1.00) - $0.62 = +$0.08 per share, meaning you expect to make 8 cents per dollar risked over the long run.

That's a good bet, even though roughly 30 of those 100 bets will lose.

Actual Outcome is what happens this one time. It's binary. You win or you lose. The outcome doesn't tell you if the probability was accurate.

This is why disciplined prediction market traders track dozens or hundreds of bets over time, not individual wins and losses. One correct call proves nothing. A pattern of +EV decisions compounding over months proves edge.

Real Example: The 2024 Election Whale

In October 2024, a French trader deposited $30 million into Polymarket across four accounts and bought "Trump wins" at an average price of 54¢. By Election Day, those contracts were worth $48 million. On November 6th, when Trump won, they were worth $55.6 million. The trader made $25.6 million.

Reddit erupted: "This guy's a genius."

Maybe. Or maybe he bought a roughly 50-50 race (the market was near 50¢ when he started) and got lucky. To know if it was skill, you'd need to see his full trading history across hundreds of bets. One win (even a $25 million win) doesn't prove edge.

Conversely, in January 2025, a trader bought $12,000 of "No" shares on "Will there be a new COVID variant designated by WHO in Q1 2025?" at 22¢. Research tracking World Health Organization designation patterns over the past decade suggested the true probability was closer to 15% (meaning the market was overpricing "Yes"). That's a +EV bet. No new variant was designated. The trader won $9,360. But if a variant had been designated, the loss wouldn't have made the bet bad, just unlucky.

The longshot bias trap

Some beginners chase longshots because "10% at 8¢ is great value!" It's only great value if the true probability is higher than 10%. If the true probability is 5%, you're paying 8¢ for a 5¢ contract. That's negative EV, even if it occasionally wins.

In February 2025, Polymarket offered: "Will Biden mention the word 'discombobulate' in his next press conference?" It traded at 9¢. Volume: $62,000. Biden had never used the word in public. A trader bought $500 at 9¢, thinking it was funny. Biden didn't say it. The market resolved "No."

The trader posted: "I knew it was a joke bet, but someone was offering 9¢!"

The fallacy: Just because someone is offering a price doesn't mean it's +EV. Markets can misprice, but you need a reason to believe your estimate is better than the crowd's.

Practical tip: Before entering a trade, write down your personal probability estimate. Then compare it to the market price. If they're within 3-5 percentage points, the bet probably isn't +EV after accounting for fees and liquidity costs. Only trade when you have a significant edge, not just a hunch.

For more on how to identify +EV opportunities, see buying a dollar at a discount.

---

5. Zero-Sum Dynamics: Why The Market Can Be Smart and Wrong Simultaneously

Every dollar won in a prediction market comes from someone else