Predictions Lab

Predictions Lab is a full-stack prediction markets platform built with a Python backend, JavaScript frontend, Firebase Authentication, and Firestore. The platform allows administrators to create prediction markets for virtually any event, while users participate using simulated currencies: SigCoin for general markets and LiveCoin for live events. Every user starts with 100 of each currency to trade.

The backend prices each market using a logarithmic market scoring rule (LMSR), an automated market maker commonly used in prediction markets. Rather than requiring a buyer and seller to be matched for every trade, LMSR ensures that shares can always be bought while keeping the market maker’s maximum possible loss bounded regardless of the outcome. Users buy shares representing different possible outcomes, with each share priced below 1 SigCoin or LiveCoin. When a market resolves, shares corresponding to the correct outcome pay out 1 coin each, while all other shares become worthless.

Live events take place on a dedicated Live Stage page, where users can watch the market update in real time as the event unfolds.

A leaderboard ranks users according to the value of their SigCoin and LiveCoin wallets, adding a competitive element to both general and live prediction markets.

Administrators can create prediction markets with two or more tradeable outcomes, select the currency used for trading, and configure an initial liquidity pool. The same panel also supports lightweight polls, where users can answer a question once in exchange for 2 SigCoin.

These polls can also be used to create experiments inspired by the Keynesian beauty contest: rather than trying to identify an objectively correct answer, participants try to predict what the group as a whole will choose. This creates opportunities to explore game theory, collective behavior, and strategic decision-making—topics that fit naturally with SIGecom’s focus on economics and computation.

Predictions Lab collects only limited user data, with its collection and use described transparently in the Privacy Policy and Terms of Service on the official project website.