Where Prediction Markets Misprice Earnings (And How to Find the Edge)
Prediction markets are the most honest price discovery mechanism on earth. No analyst incentives. No CEO spin. Just money where your mouth is.
But they still get earnings wrong. Consistently. And the gaps between what the market prices and what history tells us are where the real edge lives.
The Spread Is the Edge
Here's the simplest framework: take a company reporting earnings this week. Polymarket says there's a 60% chance of an EPS beat. Now check the historical beat rate — the company has beaten estimates 85% of the time over the last 8 quarters.
That's a 25-point spread. The market is underpricing the beat. That's your signal.
The edge isn't in knowing whether a company will beat. It's in knowing where the crowd is systematically wrong about the odds.
Why Prediction Markets Misprice
Three recurring patterns:
- Recency bias: One bad quarter tanks the market's confidence, even when the 5-year beat rate is 80%+. Prediction markets overweight the last data point.
- Narrative contamination: A CEO controversy, a product delay, negative press — these affect prediction market odds even when they have no bearing on the quarterly EPS number.
- Liquidity gaps: Smaller-cap earnings markets have thin order books. A few large trades can push odds 10-15 points in either direction. The odds reflect positioning, not information.
What I Built to Find These Gaps
Polyearnings surfaces exactly this data:
- Upcoming earnings calendar with Polymarket odds baked in
- Historical beat rates — last 4/8/12 quarters, by company
- Spread analysis — where are the biggest gaps between market odds and historical performance?
- Live order book depth — is the price driven by information or positioning?
- AI-generated edge signals — pattern recognition across the full dataset
The free tier shows the calendar and basic odds. Pro shows where the market is wrong.
A Real Example
Last quarter, a major tech company had a Polymarket beat probability of 55%. Their historical beat rate over 12 quarters was 92%. The spread was 37 points.
Why was the market so bearish? A negative earnings preview from one analyst and a CEO interview that spooked retail. Neither had anything to do with the actual EPS number.
They beat. By 12%.
The lesson: prediction markets aggregate sentiment, not just information. When you can separate the two, you find edge.
The Framework
Every earnings event, I run the same checklist:
- What does Polymarket say? (current beat probability)
- What does history say? (beat rate over 4/8/12 quarters)
- What's the spread? (anything over 15 points is interesting)
- What's the order book look like? (thin = noise, deep = signal)
- What's the narrative? (is sentiment detached from fundamentals?)
When the spread is large AND the narrative is disconnected from fundamentals — that's the play.
I built Polyearnings to automate this entire workflow. Check it at polyearnings.com — iOS app shipping this week.