We trade event contracts on our own account, and we read the same markets for the institutions exposed to the outcomes they price.
What we doMARPON Capital trades prediction markets and other event-driven markets on its own account. Systematic and discretionary strategies run on the same capital, from market making and latency-sensitive execution through to positions built on macro, corporate and regulatory research.
Over five US presidential cycles, the Iowa Electronic Markets were closer to the final result than the polls published at the same moment, 74% of the time. On the night before those elections the market missed by 1.33 points on average.
The same pattern holds on the venues trading today. Across roughly 2,200 political markets in the closing weeks of 2024, whenever the price was confident, above 80 cents or below 20, the outcome matched the price 96% of the time.
It is not only elections. When Google and Ford ran prediction markets inside their own companies, those markets beat the firms’ own expert forecasters, cutting forecast error by as much as a quarter.
The reason is money. A forecaster who is wrong loses nothing. A trader who is wrong pays for it, so every point of probability is worth arguing over, and the people arguing hardest are the ones holding a position on the answer.
A price that good is still not a decision, and it is not yet yours.
The screen price is an input, not an answer. Proprietary adjustment models recalibrate it and reduce forecast error against the raw quote, and every contract is read for what it actually settles on, since the wording can make a price mean something quite different from the risk in front of you.
Nothing on this site is an offer or invitation to invest.