Friday, March 9, 2012

2012.03.09 Weekly DJIA $ Gold $ Silver Predictions































Very short time frame charts






































The structural models for gold and silver seem to be working out well - but the DJIA structural model is still, to put it kindly, in flux.  A non-technical explanation is that while these models are adaptive up to a point - that is they can  modify their parameters with new data - the DJIA model is treating the 45 degree angle spike off the predictive curve as short term volatility because of the non-linear signature of the time series at that point.  You don't need fancy math to see that the texture of the time series at the point where the actual DJIA curve jumps off the predicted curve is very different from the rest of the curve - and this is occurring at different time scales.  Which does not mean the model is correct, it is just an interesting development.

I hope everyone has a good weekend - and if you are close to the poles, enjoy the nice Aurora Borealis/Australis!

Visit econocasts.net for free sample trading model downloads. 

3 comments:

Anonymous said...

Thank you and have a good weekend too, Paolo.
Lili

chuck said...

Hi Paolo,
Fascinating correlation in Silver /Gold. Do you also trade? What is the basic premise of the cycle construction?

Paolo said...

Hi Chuck, indeed I find the correlation between Au and Ag highly amusing. I trade based on AI models that I run, currently I am focused on VIX, XIV, TVIX and UVXY. These models are distinct from the cycle models, which provide longer-term view of the corresponding time-series.

The basic premise of cycle model construction is that there are defined business/commodity cycles of varying period and phase which when combined with pattern recognition algorithms based on non-linear parameter estimates give a long term view on the behavior of the time series in question - DJIA, Gold and Silver. Gold and Silver seem to be better predicted in the short term than the DJIA. As to why this might be - your guess is probably as good as mine.