Sunday, February 15, 2015

2015.02.13 VIX Cycle Model Chart

2015.02.13 VIX Cycle Model Chart

The VIX cycle model suggests continued upward pressure on VIX, especially given the current Z-score of 3.4. The usual caveat is that this model is useful for estimating longer term direction rather than day-to-day swings in price.  Back in August of 2014, the VIX cycle model suggested a gradual rise into 2015. The timing was more accurately predicted than the price levels. That model run is on the blog and also shown below.  The difference between the actual and predicted is an important input into a neural net model I use, since it contains the cyclical information, but there are other factors used in the neural net model that clearly influence short term behavior which cannot be captured by cyclical analysis.

2014.08.01 VIX Cycle Model Chart


Anonymous said...

It's ok. We are getting very close to the top.$SPX&p=W&yr=20&mn=0&dy=0&id=p46752639806&a=237616938

Thank you, Diego.

Anonymous said...

Many thanks for the update Paolo - did you rerun the VIX model or just update with the latest VIX values? I'm wondering if there is some value or information in cases where model reruns converge (minimal changes from last rerun).

On the topic of Z-score values, there there a threshold in terms of peak value or an integral of Z-score that alerts to when the model might not be suitable anymore? (Such as with the DJIA model)

I'm wondering at point which the model diverges from reality does one decide to reassess and cut short waiting for a reversion to the (model) mean.

I find your work intriguing and am thinking about obtaining your long-term models soon (any planned updates this year?)

Mr. Anon