Monday, November 23, 2009
Trading Models and the Scientific Method
Developing trading models is fun, and I have done similar work trying to predict the outcomes of other systems with numerous input variables, such as sport events.
However, the ultimate question for any algorithm or model or signal is whether it stands up to rigorous testing via the scientific method. Does the model or algorithm, when stripped of all hindsight and other biases, predict outcomes better than a random signal?
It is on this question that most technical analysis/modeling/algorithmic trading approaches fail.
However, the ultimate question for any algorithm or model or signal is whether it stands up to rigorous testing via the scientific method. Does the model or algorithm, when stripped of all hindsight and other biases, predict outcomes better than a random signal?
It is on this question that most technical analysis/modeling/algorithmic trading approaches fail.
# posted by MATT E @ 1:20 PM 
