Log Periodic Power Law

In mid May I posted analysis based on Didier Sornette's observation that markets undergoing super exponential growth cannot continue for long and that an approximate time to a regime change can be estimated in advance. That analysis predicted May 22 for a change. The market began a correction on that date. Didier suggests that the prediction date can be estimated months in advance within a couple of weeks. May 22 began a short pull back but none the less the super exponential growth was stopped for a while. Today it appears that the market has doubled down on more obvious super exponential growth. On a log-price-scale chart a straight upward sloping line indicates exponential growth. Super-exponential growth will show upward curvature. Here is the NASDAQ on a log-scale weekly chart:
[ATTACH=CONFIG]19411[/ATTACH]
It appears to my eye that there is indeed an upward reaching curvature.
The next chart shows a fitted Log-Periodic Power Law equation to the NASDAQ daily close values.

[ATTACH=CONFIG]19410[/ATTACH]

The date that comes from the fitting is 2.3 trading days away. I note that Pascal has suggested caution based on his observations, I guess this analysis suggesting the same thing. If you Google Didier Sornette's name you can read more about his methods.

Do I understand correctly that based on this the top will be this Wednesday morning?

[QUOTE=Chern;26212]Do I understand correctly that based on this the top will be this Wednesday morning?[/QUOTE]

That's what the analysis shows but I should mention that plus or minus a couple of weeks should be taken into account so I don't place a lot of weight on the particular day.

Mike,
Mathematically this makes sense but it seems so far that leading stocks are behaving strong. Wouldn't a breakdown of leading stocks need to occur first to bring validity to this prediction model? Also, is this model predicting an intermediate correction is coming or bear market?

[QUOTE=MTman;26214]Mike,
Mathematically this makes sense but it seems so far that leading stocks are behaving strong. Wouldn't a breakdown of leading stocks need to occur first to bring validity to this prediction model? Also, is this model predicting an intermediate correction is coming or bear market?[/QUOTE]

Not really. There was little breakdown in leading stock on 31 October 2007 market top. What Didier Sornette says about this analysis is that it attempts to catch the herding nature of man, when everyone is leaning in the same direction. Everone leaning in the same direction is unstable. It is the market version of climax activity. A stock in a climax run shows no weakness but that doesn't mean that it isn't very dangerous. This kind of analysis is new to me. And as any new technique it needs to prove itself before I feel comfortable with it.

Mike,

If you have the time/inclination to update your chart(s) here, it would be much appreciated.

Thanks,

Gary