? Markets & Resolution

Every Polymarket market is a binary question with a defined resolution date and an oracle assigned to report the outcome. Understanding how markets are created, how the CLOB market-making layer works, and how resolution cascades into settlement is the foundation of understanding the entire system.

Anatomy of a Binary Market

Every Polymarket market has five defining components. The question must be binary and unambiguous. The resolution criteria determine exactly what 'YES' means. The oracle reports the outcome. The end date is when the event is considered complete. The liquidity pool provides the trading depth.

Question
"Will ETH exceed $4,000 by June 30, 2026?"
Binary YES/NO, specific, time-bound
->
Resolution Criteria
"ETH/USD price on CoinGecko at 23:59 UTC"
Objective data source, no judgment required
->
Oracle
polymarket-oracle
Reports result after end date
->
Liquidity Pool
$2.4M TVL
AMM backing, market-maker depth
Resolution date: June 30, 2026  |  Current price: $0.62 (62% implied probability)  |  Volume (24h): $340K  |  Shares outstanding: 62M YES 38M NO

CLOB Market-Making: How Professional Liquidity Works

Polymarket's CLOB (Central Limit Order Book) layer is where professional market makers post firm bids and asks. Think of it like a limit order book on Binance or NASDAQ - but on-chain. The AMM sits behind it as the backstop. Here's how the interaction works.

1
Market maker posts a bid at $0.61

MM has a bid on the book: "Willing to buy YES shares at $0.61, up to 5,000 shares." This is a limit order - it only fills if someone is willing to sell at that price or better.

v
2
Another MM posts an ask at $0.63

Another market maker posts: "Willing to sell YES shares at $0.63, up to 8,000 shares." The spread between the best bid ($0.61) and best ask ($0.63) is now $0.02 - or about 3.3% of the midpoint.

v
3
Traders cross the spread

If a trader places a market order to buy at $0.62 (midpoint), it immediately fills against the best ask at $0.63. Better price than AMM. If they place a limit order at $0.61, it sits on the book until someone crosses.

v
4
AMM backstop: fills at current price if no book match

If the trader's limit order doesn't cross, it fills against the AMM. The AMM price at this moment is $0.62 - exactly the mid. The AMM always has liquidity; it's the guaranteed quote layer.

Market Best Bid Best Ask Spread AMM Mid 24h Volume
ETH > $4K by June $0.61 $0.63 $0.02 $0.62 $340K
Fed rate < 4% by Dec $0.44 $0.46 $0.02 $0.45 $120K
ETH merge date $0.99 $0.995 $0.005 $0.992 $2.1M

High-volume markets (like the ETH merge example above) have much tighter spreads because more market makers compete. Long-tail markets with low volume can have 5-15% spreads, making it expensive to trade small sizes.

PnL at Different Probabilities - Pick Your Entry

Choose a market's implied probability with the slider. See exactly what your profit/loss looks like if the market resolves YES or NO, at different position sizes.

Buy YES
Price per share $0.65
Stake amount $100
Shares bought 153.8
YES resolves
+$53.1
Payout = 153.8 $1.00 = $153.8
NO resolves
?$100.0
Shares worth $0
Expected value (your estimate: 80%)
+$32.5
EV = 0.80 $53.1 ? 0.20 $100
Buy NO
Price per share $0.35
Stake amount $100
Shares bought 285.7
YES resolves
?$100.0
NO shares worth $0
NO resolves
+$185.7
Payout = 285.7 $1.00 = $285.7
Expected value (your estimate: 80%)
+$128.5
EV = 0.20 (?$100) + 0.80 $185.7
PnL sensitivity to entry probability
5%25%45%65%85%95%

? Resolution Process - Step by Step

Event date passes

The market's end date arrives. No more trading happens, or trading continues but the outcome is already determined.

Oracle reports

The assigned oracle reads the data source (e.g., CoinGecko price) and submits the outcome to the smart contract. This is on-chain.

Dispute window opens (3-7 days)

Any ?U holder can challenge by posting a bond. If no one challenges within the window, the oracle's report is final.

Challenge -> ?U holder vote

If challenged, ?U holders stake on the outcome they believe is correct. Token holders with skin in the game vote with their money.

Outcome finalized, settlement begins

Winners redeem YES/NO shares at $1.00. Losers receive $0. AMM pool is drained. Market closed.

Example: A market on "Will Bitcoin exceed $100K by Dec 31, 2026?" closes on that date. On Jan 2, the oracle reads CoinGecko's BTC/USD price and reports YES. The dispute window runs 5 days. On Jan 7, with no successful challenge, the outcome is finalized. ?U holders exchange winning tokens for USDC.

Liquidity Provider Mechanics & Market-Maker Hedging

Liquidity providers in a Polymarket binary market deposit YES and NO outcome tokens into the AMM pool. They earn a spread on every trade, but they carry directional risk - the outcome risk is not hedged unless the LP actively maintains a balanced position.

LP without hedge
Deposit 60M YES + 40M NO tokens. If YES resolves, the YES tokens pay $1 and NO tokens go to $0. LP makes ~60% gross return on YES side but loses NO side. Net position depends on initial ratio and price at resolution.
->
LP with cross-venue hedge
Deposit YES tokens into Polymarket AMM. Sell equivalent YES exposure on Betfair or another prediction market to be delta-neutral. Earn Polymarket spread on the AMM, hedge removes outcome risk.
Spread earned (CLOB fills)
0.5-1.5%
per filled order
LP pool share (fee revenue)
0.5-5%
proportional to liquidity contributed
Annual revenue (on $5M pool)
$25K-$250K
at 0.5-1% daily volume
LP Risk: A pure LP position (no hedge) is not delta-neutral. If you deposit 60M YES + 40M NO into a pool priced at 60%, you are effectively long the outcome. If the market resolves NO, you lose the NO side entirely. The AMM provides liquidity, not neutrality - LPs must actively hedge if they want to earn spread without directional exposure.

Binary question design

Every Polymarket market is a binary outcome - YES or NO. This simplicity is a design choice, not a limitation. Binary markets are easier to price, easier to resolve unambiguously, and the AMM math produces a single probability figure that maps cleanly to the market's implied probability. Multi-outcome markets exist on other platforms (like Augur's old v1 design), but they introduce resolution ambiguity - what if the event partially occurs? Binary cleanly sidesteps that problem.

A well-formed Polymarket question has three constraints: (1) Objectivity - the resolution criteria must reference an objective data source (a price, a score, a date, a defined event). "Who will win the election?" has a clear answer if you specify the source. "Will it be a good year?" does not. (2) Exclusivity - exactly one outcome should be true. If both YES and NO could simultaneously be true, or if there's a third state, the market is malformed. (3) Time-boundedness - the event must have a defined end point. Markets without end dates can never resolve and become dead capital.

Resolution criteria

The resolution criteria are the contract between the market creator and the oracle. They specify the exact data source and timestamp that determines the outcome. Good criteria are specific and verifiable: "CoinGecko BTC/USD closing price on 2026-12-31 at 23:59 UTC > $100,000" is unambiguous. Vague criteria like "Does the protocol achieve its roadmap milestone?" invite disputes and undermine confidence in the market.

In practice, most Polymarket markets resolve cleanly because the data source is public and objective. Cryptocurrency prices, sports scores, and economic data are the most common categories - precisely because they're independently verifiable. Political markets use defined sources like the Associated Press or official government reports. Markets on subjective questions (like "Will this bill pass?") resolve through oracles who commit to using specific official sources before the market opens.

CLOB-AMM interaction

The CLOB layer is populated by professional market makers who post tight bid-ask spreads on active markets. Their competitive limit orders define the best bid and best ask. The AMM sits behind this - if a trader's limit order doesn't cross (fill against a counterparty on the CLOB), the AMM fills the order at its current price. The AMM price is typically very close to the CLOB midpoint, because arbitrageurs constantly buy on the CLOB and sell on the AMM (or vice versa) whenever the price diverges.

The spread between the CLOB best bid and best ask is an important signal: tight spreads mean competitive market making and good liquidity. Wide spreads indicate low competition - perhaps because the market is too small for professional MMs, too controversial, or too hard to hedge. The AMM backstop means even the widest-spread markets still have a price and can still be traded, but at a higher cost.

Market maker hedging

Professional market makers on Polymarket typically hedge their positions on external venues. A market maker holding YES shares in the ETH > $4K market will sell YES exposure on Betfair or Deribit, or open a short ETH perp position, to flatten their directional exposure. The spread they earn on Polymarket trades is their gross revenue. After accounting for hedging costs, exchange fees, and inventory risk, the net margin for a professional MM in a large Polymarket market is typically 0.2-0.8% per trade.

This cross-venue arbitrage activity is what keeps Polymarket prices connected to the broader prediction market ecosystem. When Polymarket's implied probability for an event diverges significantly from Betfair or another prediction exchange, arbitrageurs are financially motivated to close the gap. The result is that Polymarket prices reflect the best information available across all connected markets - which is precisely the point of a prediction market.

Frequently asked questions

What makes a Polymarket question 'well-formed'?
A valid Polymarket question has three properties: (1) binary - resolvable to exactly YES or NO with no third state, (2) specific - the outcome is objectively determinable by an oracle or human adjudicator, not a matter of opinion, and (3) time-bounded - the resolution date is defined. Ambiguous questions get flagged by the community or rejected by the oracle before going live. For example: 'Will Bitcoin close above $100,000 on December 31, 2026?' is well-formed. 'Will Bitcoin go up?' is not.
How does a market resolve if the event is ambiguous?
Polymarket has a multi-step resolution process. First, the assigned oracle reports the outcome based on the defined criteria. Then begins a dispute window where any ?U (conditional token) holder can challenge by posting a bond (typically $100-$250). If challenged, ?U holders vote by staking on the outcome they believe is correct. The bond system means false challenges are expensive. If the dispute succeeds, the oracle is slashed and the correct outcome is used. This mechanism is borrowed from UMA's optimistic oracle and makes falsifying outcomes economically punishing.
Can market makers hedge their positions on other venues?
Yes - and this is a key part of why Polymarket's CLOB-AMM hybrid works. A market maker who holds a large YES position in a Polymarket market can sell YES exposure on Betfair, or short the equivalent in a synthetic to offset directional risk. The spread they earn on Polymarket is their profit margin; the hedge makes it market-neutral. This cross-venue arbitrage is also what keeps Polymarket prices close to external odds - if Polymarket's implied probability diverges from Betfair's, arbitrageurs buy the cheaper side and hedge on the other venue, converging the two prices.
What is the difference between a binary market and a multi-outcome market?
Polymarket only supports binary YES/NO markets. Each market creates exactly two outcome tokens: YES shares that pay $1 if the market resolves YES, and NO shares that pay $1 if the market resolves NO. This is different from multi-outcome markets (like Augur's old design or Polymarket's predecessor Gnosis Exchange), where you could have N different settlement prices. Binary markets are simpler to reason about, settle cleanly, and the AMM math is more elegant - P(YES) is a single number between 0 and 1.
How does resolution timing affect the market?
Resolution happens after the event date, and some markets take days or weeks to fully resolve - especially when disputes arise. During the dispute window, the ?U tokens cannot be redeemed, which locks capital. Traders who want to exit before resolution can sell their shares on the secondary market (the AMM or CLOB), but only at the current market price. The longer the resolution uncertainty, the more the market discounts for that risk - which is why markets with ambiguous or disputed outcomes tend to have wider spreads and less confident pricing.
Who decides the outcome if there's a genuine dispute about facts?
Polymarket uses an 'oracle council' model - selected oracles (often the original creator or a trusted third party) make the initial report, and the community can challenge within a window (typically 3-7 days). In practice, high-profile markets with large open interest attract professional dispute participants who have the technical knowledge and capital to post a bond and argue a case. In most markets, the oracle's report goes unchallenged. Only markets with genuine ambiguity or bad-faith reporting see disputes.