How to Read Polymarket Odds: Probability, Edge & When the Market Is Wrong

Prices Are Probabilities

Every Polymarket market is binary: YES or NO. A YES share costs between $0.01 and $0.99 and pays out exactly $1.00 if YES occurs and $0.00 if not. The price at any moment is the market's consensus probability estimate.

• YES at $0.75 = 75% probability that YES occurs
• NO at $0.40 = 40% probability that NO occurs (i.e. 60% that YES occurs)
• YES at $0.75 + NO at $0.25 = $1.00 (sum always equals $1.00 on efficient markets)

This is fundamentally different from sportsbook odds. A sportsbook at -300/+250 is expressing a margin-adjusted probability with built-in vig. Polymarket prices have no vig built into the price itself — fees are charged separately on winnings. The price you see is the pure market consensus.

How Accurate Are Polymarket Odds?

Polymarket publishes a Brier score — the standard statistical measure of probability forecast accuracy — of 0.0843 across all resolved markets. A Brier score of 0 means perfect accuracy. A score of 0.25 means the forecast is no better than random for binary outcomes. Polymarket's 0.0843 is excellent calibration — in the same range as top-tier professional forecasters.

For the most liquid markets — those with 5,000+ daily contracts — calibration accuracy reaches 92-95%. When Polymarket prices a YES outcome at 70%, the event actually occurs roughly 70% of the time. This makes Polymarket odds more accurate than most individual analysts, most polls, and most pundits.

This accuracy is not accidental. Prediction markets work because they aggregate the views of thousands of participants who commit real money. The price represents the marginal trader — the one willing to put capital on the line at that specific level — which is a stronger signal than what an anonymous poll respondent says.

When the Market Gets It Wrong

Despite Polymarket's excellent overall calibration, individual markets can and do diverge from true probability. These divergences are where the edge lives. The key is understanding the common causes of mispricing:

• News overreaction: A breaking news story enters the market in seconds but the full context takes hours or days to absorb. The initial 20-cent move on a headline is often the largest and most likely to overshoot the true information impact.
• Thin liquidity at extreme prices: Markets near $0.10 or $0.90 often have wide spreads and shallow books. A single $5,000 order can move the price by 5-10 cents, creating temporary mispricings that do not reflect genuine consensus. Always check order book depth before treating an extreme price as a signal.
• Anchoring: Markets that have traded at one level for a long time can be slow to update when the reality changes. A market that sat at 80% for three months might only drop to 60% after a significant negative event because traders anchor to the old price.
• Non-informational trading: Large traders sometimes enter positions for reasons unrelated to their probability estimate — hedging, liquidity, or experimentation. A whale buying $50K of YES at 20% does not necessarily mean they think the true probability is higher than 20%.

Finding Edge: The Divergence Method

The most systematic approach to finding edge on Polymarket is the divergence method: compare the market's implied probability against your own researched estimate (or an external forecast aggregator) and trade only when the difference exceeds a threshold.

The threshold matters. Small divergences (1-3 percentage points) are noise — the market is almost certainly more accurate than you are in that range. The consensus among professional Polymarket traders is that a 5-10 percentage point divergence is the minimum threshold for a tradeable edge. Below that, the spread cost and fee structure eat any theoretical edge.

Reliable external sources for comparison:

• Metaculus: Aggregates predictions from a community of highly calibrated forecasters. Strong for science, tech, and geopolitical questions.
• Polymarket itself (historical): How similar events have resolved historically provides a base-rate anchor. If five similar proposals in the past passed at a 40% rate, a new one at 80% is worth investigating.
• Expert consensus forecasts: For political markets, election model aggregators (like base-rate analyses). For economic markets, Bloomberg or Reuters economist surveys.

The YES + NO Constraint: Checking for Pure Arbitrage

Every binary market on Polymarket has the mathematical property that YES price + NO price must equal $1.00 at efficient equilibrium. When the sum is below $1.00, buying both sides guarantees profit at resolution. When it is above $1.00, selling both sides guarantees profit.

This is the simplest edge check in existence. Before entering any directional position, check the sum. If it is materially below $1.00 (say, $0.92), your edge from a directional buy might be lower than you think — the market is not pricing the outcomes accurately on either side.

Polysharp's arbitrage detection engine monitors this constraint across all markets in real time, executing arbitrage trades automatically when the sum drops below $0.98. This provides a baseline return that funds the rest of the strategy portfolio.

The Trader's Mindset: Being Wrong About the Market vs the Market Being Wrong

The most important mental model for reading Polymarket odds is humility before the crowd's estimate. Polymarket's overall Brier score of 0.0843 means the crowd is remarkably well-calibrated. Your probability estimate, unless backed by genuinely superior data or analysis, is almost certainly less accurate than the market's.

This does not mean the market is never wrong. It means you should treat every potential divergence with a default assumption that you are misreading something. The question to ask is not "is the market wrong?" but "do I have information or analysis that is structurally better than what the aggregate of 5,000+ traders has access to?" If the answer is genuinely yes, the divergence is your edge. If the answer is "I just have a strong feeling," the market is almost certainly right and you are not the exception.

Frequently Asked Questions

What is a Brier score and why does it matter? The Brier score measures the accuracy of probability forecasts. A score of 0 means perfect accuracy; 0.25 is no better than random; Polymarket's 0.0843 is excellent. It matters because it tells you how much trust to place in market prices as probability estimates.

Should I always trust Polymarket odds over my own analysis? For markets you do not follow closely, yes. The market is almost certainly more accurate. For markets where you have genuinely non-public information or a structural analytical edge, a divergence of 5-10 percentage points is a reasonable threshold to act on.

How often are Polymarket odds significantly wrong? For liquid markets (5,000+ daily contracts), significant mispricings are rare and brief — typically correcting within minutes to hours. For illiquid markets, mispricings can persist for days because there is no active arbitrage to correct them.