- View all blogs
- Current outlook: hurricane prediction market
Current outlook: hurricane prediction market
Mark Roulston •
Earlier this month Hivemind joined Cohort-5 of Lloyd’s Lab programme to promote InsurTech startups.
As part of our activities in this programme, we are running a prediction market for the number of hurricanes that will form in the Atlantic this year, and one for the number of these hurricanes that will make landfall in the U.S. These prediction markets are now running on Hivemind’s AGORA platform which provides a mechanism to aggregate the judgments of each participating expert. All participants in the market were selected for their expertise in areas such as meteorology, climatology and statistics.
What's happened in the market so far?
The 2020 hurricane season started as one of the most active in recent history, with a record nine named storms (two of which became hurricanes) developing before the end of July. So far, the AGORA prediction market concurs with seasonal forecasts issued in early August, suggesting that we can expect nine hurricanes before the end of November. It also suggests there is a 20% chance this season could see as many as 12 or more hurricanes.
A parallel market for the number of landfalls suggests the most likely scenario is for five hurricanes to make landfall along the U.S. coast This would be more than twice the annual average since 1900 and would be especially unwelcome in a year that has already seen Hurricane Laura cause devastation when it came ashore in Louisiana as a category 4.
We'll provide a full overview of the market predictions and the views of its top three participants when the market closes on November 30th.
AGORA is a tool that can turn the views of multiple experts into unified probabilistic forecasts. It also allows these forecasts to evolve as new information becomes available to participants. It provides a structured way to harness expert knowledge, whether it exists inside or outside an organisation.
If you are interested in using AGORA to improve the information on which your decisions depend contact firstname.lastname@example.org