Sign up with Google
David Shaw: The Quantitative Edge
David Shaw was one of the most brilliant mathematicians, physicists, and computer scientists. His purpose was to combine his quantitative skills to consistently extract profits from the world's financial markets.
He seeks to profit strictly from pricing discrepancies among different securities, avoiding risks associated with directional moves in the market. His trading approach requires highly complex mathematical models, vast computer power, constant monitoring of worldwide markets by a staff of traders, and near-instantaneous, extreme low-cost trade executions, and is beyond the reach of the ordinary investor.
One concept that came up in this interview was the idea that a market pattern (or inefficiencies) may not be profitable when evaluated individually. However, it can still prove to be a profitable strategy when combined with other patterns.
Although Shaw ridicules chart patterns and technical indicators, a similar idea holds good that a combination of indicators might be a useful model to trade, even when the individual indicator is worthless when used alone.
This effect would apply to fundamental inputs as well. For example, a researcher might check ten different fundamental factors and finds that none can be used as price indicators. This does not imply that these inputs should be dismissed. Even though the factors don't provide a meaningful prediction as an individual factor, it is very well possible that the combinations of these inputs prove to be a useful price indicator.
Another important principle is that of an appropriate methodology in testing trading ideas. If a trader tries to develop a systematic approach, then he should avoid data mining through computers. This means it should not let the computer run through millions of inputs to give out profitable patterns. He believes such patterns have no predictive power because patterns can even be found in random data.
He avoids this problem of data mining by running a theoretical hypothesis before each computer test and by using rigorous statistical measures to evaluate the importance of the results.
