Day Trading Online Home Page  

The Time Series Forecast indicator shows any statistical trend in a stock's price over a time period of length 'n' using linear regression analysis techniques. Interestingly, the Time Series Forecast generates the last point of several simultaneous linear regression trends. The resulting Time Series Forecast indicator is also called the 'moving linear regression' indicator and sometimes the "regression oscillator". Day trading masters may find a use for this indicator in order to give themselves a 'flavour' of what tomorrow's action mightbe, although most prefer established techniques such as volatility analysis.

The forecast is traded like any other moving average, but the fact that multiple Time Series are used gives it a couple of advantages over 'ordinary' MAs, principly the lack of a delay when prices change rapidly. This, of course, is because the Time Series Forecast 'fits' itself to the underlying price data instead of averaging them, making it more responsive to price changes. Basically, if the current trend remains in place, the Time Series Forecast is a forecast of the next period's price level.

To calculate the Time Series Forecast you have to use a "least squares fit" technique to calculate a linear regression trendline, which attempts to fit a trendline to the price data by minimizing the distance between the price points and the linear regression trendline itself.

previous next