Volatility forecasting

[Gambit]

There is a fairly active and "lively" discussion about selling vol on ET and I thought I would extend the conversation to some other relevant topics, mainly forecasting vol.

I'll start of by throwing out some links on using the GARCH model to create a vol forecast:

https://www.mathworks.com/matlabcen...g--simulation-and-value-at-risk-applications-

https://cran.r-project.org/web/views/Finance.html

 

[Gambit]

GARCH basics explained:

 

[Robert Morse]

This is all above my paygrade but I know that NYU has done a lot of work on this.
https://vlab.stern.nyu.edu/doc/2?topic=mdls

 

[Gambit]

Bob, it's above my paygrade today too . Thanks.

 

[Gambit]

Bob, that's a great link. I had no idea there was an actual "volatility institute".

 

[stepandfetchit]

When I look at that stuff, my eyes glaze over then I nod off! Above my pay-grade as well.
However, it seems some of us lay-people may fairly safely trade volatility by just looking in the rear-view mirror after big events! (regression to the mean, works better when you trail an extreme). -- Has nothing to do with predicting when a new volatility increase will occur, however.

 

[Gambit]

When I look at that stuff, my eyes glaze over then I nod off! Above my pay-grade as well.
However, it seems some of us lay-people may fairly safely trade volatility by just looking in the rear-view mirror after big events! (regression to the mean, works better when you trail an extreme). -- Has nothing to do with predicting when a new volatility increase will occur, however.

I love that rhyme, "regression to the mean, works better when you trail an extreme." Based on my observations, that works better after a scheduled event like earnings. If the earnings announcement occurs, implied vol drops and stat vol increases for a while and then decays. However, with indexes, a high implied vol could last for a longer period of time and it is hard to predict with any certainty the end of the event. Anybody have any thoughts?

 

[Gambit]

http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.491.2514&rep=rep1&type=pdf

Here is a paper that offers immediate value. Due to my lack of mathematical sophistication, I would probably just follow the author's conclusions without critique. But even a 50/50 forecast is better than nothing.

 

[Gambit]

Here is a free introductory text on forecasting: http://www.ssc.upenn.edu/~fdiebold/Teaching221/Forecasting.pdf

I think this text and Euan Sinclair's book are a good start.

 

[stepandfetchit]

Gambit: Thanks for those links!