Reuters’ review of Renaissance Tech founder Jim Simons’ biography highlighted the importance of math skills in running a hedge fund that produced average returns of 68% a year. Simons never took a finance class. All he needed was his skill with numbers.

Anna Szymanski filed this report for Reuters:

Simons views the market as a code to be cracked, and no codebreaker wants to signal his presence. This focus explains Renaissance’s hiring practices. It mostly favors mathematicians over traders or economists and makes employees sign super-strict non-disclosure agreements. It also explains Simons’ focus on collecting and scrubbing enormous amounts of data in the search for statistically significant patterns.

This is where the book gets interesting. Simons views gyrations in financial markets as a hidden Markov model: a sequence of seemingly random events that are actually governed by concealed variables. He believes that markets exist in different states, and that short-term relative movements in asset prices can be better predicted by using algorithms that infer this hidden structure. While he originally incorporated some gut instinct into his trading, he eventually gave up most control to the algorithms.

Importantly, a model built using this concept isn’t designed for perfection. Instead, it’s designed to make better predictions. Understanding the difference between the probability of success and the guarantee of success appears to have helped Renaissance survive the fate of rivals like Long-Term Capital Management, which failed after Russia devalued the ruble in 1998.

The description of how the Medallion trading system functions is downright beautiful, even for readers who aren’t math geeks. It designs a multitude of trades that work in concert to generate high returns at low risk across asset classes, in such a way that patterns normally remain hidden to other traders. This mathematical ballet produces shockingly high Sharpe ratios – a common measure of risk-adjusted returns. In early 2003 the Medallion fund achieved a ratio of six, nearly twice what the biggest quant funds would hope for.

Medallion had some help, particularly from the use of basket options. These are derivatives whose value is tied to a basket of underlying securities. Banks like Deutsche Bank and Barclays held and traded the actual securities according to instructions from Renaissance computers. This enabled the firm to wager far more capital than it controlled, seriously boosting returns. But unlike more conventional leverage strategies, such as those used by LTCM, the risk remained mostly with the banks. The most Renaissance could lose was the option premium and its collateral.