If it could be ,then how to predict it?
Thank you
Thank you
Could volatility be predicted ?
- Thread starter flyingforget
- Start date
YES it can be predicted.
However, the direction of price to the volatility...
NO.
How???
Not worthwhile sharing it here at ET.
Mark
Volatility might be defined simply as the daily gyrations of a given market (eg CL,ES,YM). Therefore you can do better than 'prediction', you can establish the datasets of the metrics and parameters of that market.
As I have a professional background in futures, I have for some five years at ET pointed to where the money is in a market, which is in the gyrations. Therefore you can take it as proven that very few, if any, coming through ET will ever want to make themselves rich because they do any manner of things other than adopt a methodology to try and take the money where it arises in a market, open through to close, day in and day out.
As I have a professional background in futures, I have for some five years at ET pointed to where the money is in a market, which is in the gyrations. Therefore you can take it as proven that very few, if any, coming through ET will ever want to make themselves rich because they do any manner of things other than adopt a methodology to try and take the money where it arises in a market, open through to close, day in and day out.
anything can be predicted with a certain margin of error.
But, the volitility does not follow a normal distribution, and the first step is properly determening / predicting Volatility is determining which distribution it follows.
Once you have that, you can begin to model.
If you use a Gaussian distribution - you *may* see good results for a bit, but the outliers will screw you. And, in reality, these aren't really outliers - they are part of the curve.
Not an easy task.
But, the volitility does not follow a normal distribution, and the first step is properly determening / predicting Volatility is determining which distribution it follows.
Once you have that, you can begin to model.
If you use a Gaussian distribution - you *may* see good results for a bit, but the outliers will screw you. And, in reality, these aren't really outliers - they are part of the curve.
Not an easy task.
Yes, it can.