Short Term Trading Member Forum

[techcharter]

Mean Model is based on Regression Analysis

... but is not a linear regression analysis. I use SSE (sum squared error) to fit the set of curves to the data. I think of curve fitting rather than regression analysis since I have an engineer's point of view, not a mathematician's. I use correlation to validate how reliable the Mean Model and prediction curves fit market data. And x out of y event statistics on the Mean Model for determining market turn points. This is a complex algorithm which uses several mathematical tools and engineering concepts.

I think most people using statisics get in trouble when they try to take raw market data and count how many conditional events occur out of the total data set. Stochastics can give an overbought or oversold signal, yet such a signal can persist quite a while before the market finally makes the turn.