Posted By: john   Date: Thu Jul 16, 2009

Title: past revisions of cycle chart

Message:
  just last week the cycle chart looked like there could be a 10% more or less drop from the current DJI to get to the red line or projected path of the DJI. You revised the projected chart. There still was a down turn but more gradual and only about a 500 point drop from the current number.
 
  Yesterday, the DJ took off to the upside.
 
  My question is- if in the past where there may have been revision in the cycle chart- were there times that it looked like last week. The projected chart last week was showing a large gap between actual and projected trend in the next 10 days. The Chart was revised according to your studies and data. Then the DJI takes off in the opposite direction that was projected after the change? Then you have to revise your charts again with the new data.
 
 

Response:
  The cycle model does not have a learning algorithm like our standard model so shocks are taken into account by recalculating the cycle parameters. It is impossible to know a priori what effect new parameters will have on the predicted trend. As I state on the forecast page, this model is experimental. So far, I don't find that it is very useful in predicting the market direction. I plan on keeping it for another month and I may then replace it with something else that I have been working on that may do a better job.