digest
Short Summary
Main recommendation: Keep investing in index funds rather than trying to replicate the Medallion Fund, because its exceptional returns are real but effectively inaccessible to outside investors. The video argues that Renaissance Technologies' Medallion Fund challenges market efficiency as a model, but its edge comes from proprietary data, quant models, and talent that ordinary investors cannot access.
Key Points
- The Medallion Fund reportedly returned about 66% per year on average before fees from 1988 to 2018, which raises questions about whether markets are efficient and whether index funds still make sense.
- Jim Simons, a mathematician and code-breaker, founded Renaissance Technologies and recruited his old code-breaking colleague Leonard Baum, who had developed the Baum-Welch Algorithm for hidden Markov Models.
- Renaissance was among the first quant investors, building massive cleaned historical datasets and using algorithms that found patterns without needing to know why they existed.
- A key breakthrough came in the early 1990s when Robert Mercer and Peter Brown, from IBM's speech recognition group, helped the firm conquer stock trading in 1995; Mercer noted their trades won only 50.75% of the time, which is enough when trading millions of times with leverage.
- Medallion has been closed to outside investors since 2003 and caps its size at about $10 billion by distributing profits annually, so most of its gains were unavailable to the public.
Details Worth Noting
- Concrete instructions: The video gives no step-by-step method for replicating Medallion. Its only actionable guidance is for the investor: before making any active trade, ask who is on the other side and what they know that you don't; and, given that Medallion is inaccessible, continue holding index funds.
- Bradford Cornell's 2019 paper "Medallion Fund: The Ultimate Counterexample" found that $100 invested in Medallion in 1988 would have grown to about $398.7 million by 2018, with no negative return over that period.
- Cornell's perfect-foresight market timing model, using monthly returns, grew $100 to about $331,288 over the same period, still less than 10% of Medallion's result.
- The fund's returns were not explained by known risk premia, according to Cornell, and the employees are bound by legally binding secrecy, so the exact methods are unknown.
- The video describes the paradox of skill: when equally skilled investors compete for alpha, luck determines the winner, and Medallion's future edge is uncertain given rising publicity and competition from other quant investors.
Bottom Line
Medallion's record is a genuine challenge to market efficiency, but it does not translate into a practical strategy for most investors, so the reasonable choice remains low-cost index investing.
