Why does prediction market consensus beat single-market prices?

Consensus beats single-market prices through three mechanisms: (1) wisdom of crowds — aggregating diverse participant information sets; (2) vig removal — normalizing prices to true implied probability; (3) noise reduction — downweighting thin markets by volume.

Statistical Analysis

Why Consensus Beats Single-Market Prices

This analysis explains why aggregated consensus from multiple prediction markets provides a more reliable probability estimate than data from any single platform, with supporting evidence from calibration research.

Updated March 2026 · 7 min read

Prediction market consensus aggregates prices from 25+ regulated sources and volume-weights them into one probability. It consistently perform better thans any single market price by 6–10 percentage points in accuracy because it pools information from more participants, removes vig, and is resilient to thin-market noise. This is the same principle behind superforecasting and ensemble models in machine learning.

The Numbers

65%

Accuracy — manual cherry-picked source

68%

Accuracy — single market devigged

74%

Accuracy — Meridian consensus (25+ sources)

Across 2,000+ resolved events in our backtest, consensus predictions reduced prediction error (Brier score) by 18% compared to the best available single-market price.

Why Aggregation Wins: 3 Mechanisms

1. Wisdom of Crowds

Each prediction market represents a different set of participants with different information sets. Aggregating across markets pools that collective intelligence. A participant with private knowledge in Market A moves prices there — and the consensus captures that signal even if Markets B and C haven't priced it in yet.

2. Vig Removal

Every prediction market charges a spread between YES and NO prices. A $0.63 YES and $0.42 NO doesn't mean 63% probability — the combined price is $1.05, inflated by a 5% vig. Meridian Edge deviggs each source before aggregating, returning true implied probabilities.

Raw market price (with vig)

63%

Devigged single market

60%

Meridian consensus

58%

Example: three sources all show slightly different prices for the same event. Consensus aggregates them into one calibrated probability.

3. Noise Reduction

Thin markets — those with few participants and low volume — are more susceptible to noise. A single large order can move a thin market significantly. Consensus downweights thin sources by volume, so a single outlier can't distort the aggregate probability.

Divergence: When Markets Disagree

One of the most valuable features of consensus is divergence detection — identifying when different prediction markets price the same event significantly differently.

Example divergence — live data format

Will the Fed cut rates before June 2026?

Market A67%

Market B59%

Market C62%

Meridian Consensus62.3%

The 8-point gap between Market A and Market B is flagged as a divergence. Meridian Edge returns these via the /opportunities endpoint for further investigation.

The superforecasting parallel

Philip Tetlock's superforecasting research found that aggregated forecasts from diverse groups consistently perform better than individual experts. Prediction market consensus applies the same principle — but with real money as a weighting mechanism, which selects for calibration over time.

When Single-Market Prices Are Sufficient

There are cases where a single market price is acceptable:

When only one market lists the event (no other source to aggregate)
When the market has extremely high volume and tight spread (already well-calibrated)
When you need order book depth for execution purposes (aggregation doesn't help here)

For analysis, dashboards, AI agents, and research, consensus is almost always the better choice.

Access Live Consensus

Meridian Edge consensus is available via API and dashboard. The live dashboard shows real-time consensus for all active events. The API returns JSON with probability, confidence, source count, and spread.

Frequently Asked Questions

How much more accurate is prediction market consensus than a single market?

In backtests across 2,000+ resolved events, Meridian Edge consensus achieved ~74% accuracy versus ~68% for the best single-market devigged price. This represents an 18% reduction in Brier score (prediction error). The gap is larger for events with thin markets or high cross-market divergence.

Why does removing the vig improve accuracy?

The vig (overround) is the profit margin built into prediction market prices. A $0.63 YES and $0.42 NO sums to $1.05, inflating both prices above their true implied probability. Removing the vig brings each price back to its true implied probability before aggregation.

What is a prediction market divergence?

A divergence occurs when multiple prediction markets show significantly different prices for the same event — typically > 5 percentage points. Divergences indicate disagreement between participant groups and are often resolved as more information becomes public. Meridian Edge flags all divergences above 5¢ via the /opportunities endpoint.

Does consensus work for all event types?

Consensus is most effective for events with multiple active markets and high participation. Sports (NBA, NHL, NFL) and major economic events typically have the strongest consensus signal. Niche events listed on only one platform use the single-source price.

See consensus vs single-market live

Live dashboard. No signup. All active events.

Open Live Dashboard

Or explore the API

For informational purposes only. Not investment advice. Risk Disclosure