Ax, Straus the Data Collector and Axcom
Ax wasn’t a likeable character. He got into a lot of arguments/fights. His personal life was in disarray. His early research wasn’t original either - it was comparable to that of trenders.
Straus was a data collector. The Telerate machines at trading floors didn’t have an interface for investors to collect and analyze information (a laid-off Michael Bloomberg would later change that). Straus collected tick data on commodities, futures and stocks. This was above and beyond what Straus was hired for.
Straus foraged and cleaned data that the rest of the world cared little about. He looked for unusual spikes, dips and gaps.
Ax and Straus moved to the West Coast, forming Axcom. Simons received 25% of the profits and agreed to provide trading help and communicate with the new firm’s clients.
Carmona Brings in the Stochastic Guns
Axcom needed new ways to make money. Brazilian weather history didn’t help predict coffee prices. Public sentiments and holdings of other futures traders didn’t help much either.
Simons, Ax and Straus didn’t believe that the market was truly a random walk. A probability distribution could capture future prices as well as any other stochastic process.
Carmona (Irvine -> Princeton) was brought in for his knowledge of stochastic processes. Linear regressions did not do shit despite more data. Carmona started experimenting with kernel methods. Straus was running out of data - he even started to model data.
In Long Island, Laufer was also using kernel methods to analyze patterns in data sets. He was set to share with Simons.
Simons wasn’t a fan of the kernels. They couldn’t reduce them to a set of standard equations.