On 22 July 2013, Émile Servan-Schreiber announced that he had just launched the prediction market for the Good Judgment Project’s third season on a new platform. Two updates, posted 51 seconds apart, carry the same announcement with only a punctuation change. The message marked one launch within the project’s continuing forecasting work.
Where Lumenogic fitted into the project
Hypermind’s own company history places the work in its Lumenogic period. NewsFutures had been rebranded as Lumenogic in 2010. In 2012, according to that history, Philip Tetlock invited the company into the IARPA-funded Good Judgment Project to contribute prediction-market expertise. The July 2013 update is a contemporary milestone within that collaboration.
IARPA’s official account of its Aggregative Contingent Estimation programme describes the research objective: improve intelligence forecasting by eliciting probabilistic judgements, weighting and combining them, and testing the resulting forecasts against real events. A market was one way to obtain and aggregate that information. It was part of a research effort comparing methods, rather than simply a public betting launch.
What later primary research shows
A subsequent paper, Distilling the Wisdom of Crowds: Prediction Markets vs. Prediction Polls, names Servan-Schreiber among its authors with a Lumenogic affiliation. Published online in April 2016, it reports an experiment involving more than 2,400 participants, 261 events and two tournament seasons.
Participants were randomly assigned to markets or prediction polls. Market prices outperformed a simple average of poll forecasts in both seasons. But team polls did better when their forecasts were combined with statistical methods accounting for recency, past performance and recalibration. The result supports comparing complete forecasting processes, not declaring one mechanism universally superior.
The later study provides methodological context for the earlier launch; it does not establish the third-season platform’s results. The infrastructure announcement nevertheless identifies a concrete step in organising collective forecasting: providing a market through which participants could contribute to a shared assessment. The journal’s history of proven forecasters and Superforecasting follows the broader research story. In July 2013, Servan-Schreiber’s message focused on the platform needed to support another season of that work.
