How Predict Cornell Works

This page gives the mechanics without heavy finance jargon. If you can read a chart and compare percentages, you can use this app well.

About the Project

Predict Cornell is a campus prediction market where students trade on yes/no questions. The live percentage is a crowd estimate of how likely YES is right now.

How Trading Works

  1. Pick YES or NO.
  2. Enter how much you want to risk.
  3. You receive shares for that side.
  4. When the market resolves, winning-side shares pay out.

Probability as a Crowd Estimate

If a market is at 70%, that means traders collectively price YES as more likely than NO. It does not guarantee a result. It is a live estimate, updated by real trades.

In short: probability here is a "market belief" number, not an official prediction from Cornell.

Shares and Payouts

Shares are your position size. More shares means a bigger stake.

Example: You spend $30 on YES and receive 41.2 YES shares.

  • If resolved YES: those 41.2 shares pay out.
  • If resolved NO: YES shares do not pay out.

Why Price Moves

When someone buys YES, the system adds YES exposure and pushes YES probability up. When someone buys NO, YES probability moves down.

Bigger trades move price more than small trades. This is intentional so the chart reflects conviction.

LMSR in Plain Language

The app uses LMSR (Logarithmic Market Scoring Rule), a pricing formula designed for prediction markets. It keeps prices smooth and always gives a quote for both sides.

A liquidity parameter b controls sensitivity: higher b means prices move less per trade, lower b means prices react faster.

Read the original paper: Hanson (2007) on LMSR.

← Back to markets