This morning, I initiated a new trading system on collective2 called functional analysis. This is an automated system that uses a proprietary method to select markets and generate trading signals. The portfolio used by this system will be composed of futures contracts and the time frame for trading will be on the order of days to weeks (a swing system time frame). When trades are entered, they are accompanied by a stop and target. Position sizing, risk management, stops and targets are all volatility based. The initial portfolio is:
long: silver, orange juice, corn, soybean meal, and palladium
short: emini russell 2000, crude oil (wti), natural gas, cotton, and coffee
The title of the system is due to the use of functional analysis in the method used to select which markets to trade. A concise definition from Wolfram MathWorld is: “Functional analysis is a branch of mathematics concerned with infinite-dimensional vector spaces (mainly function spaces) and mappings between them. The spaces may be of different, and possibly infinite, dimensions. These mappings are called operators or, if the range is on the real line or in the complex plane, functionals.”
There are many ways to use advanced mathematics in trading systems. The functional analysis system is based on viewing time series of asset prices as functions. If one considers such time series, they can be treated as discrete functions. The domain can be time or a dimensionless index. The range is the historical price. By considering time series as functions, an enormous amount of mathematics is available for exploitation. However, if one is able to transform the discrete function into a continuous one, then an even larger amount of mathematical knowledge is available.
Thus, by viewing markets as functions, I was able to utilize some methods from functional analysis to develop this trading system.