Thanks for the ref. I've put it on my wish-list... is it only applicable to options or can it be interpreted more generally?
Quantifying randomness: variance ratio
- Thread starter stephencrowley
- Start date
Well, if the Runs Test works then the Variance Ratio testw ill definately work as the VR test is probably much more academic in nature.. of course you can always draw lines on some random walk and call that a "trend" but these tests help you determin if you are drawing "real" trend line or are just being fooled by randomness.
If you are not interested in options I wouldn't recommend buying the book.
Out of curriocity I made a quick calculation using the Parkinson number in a sample of DJI data and also in a sample of a strong uptrending index.
DJI gave an almost random data series with a ratio 1.73 (1.67 = random)
The uptrending market gave a ratio of 0.89.
Here is a variance ratio profile for 30 days.. shows that the dow is slightly mean-reverting and the FTSE is slightly trending and then completely random after 20 days or so.
Of course this is a small sample so take it with a grain of salt.. there are some confidence bounds around those numbers but i dont have the calculates handy right now.
Of course this is a small sample so take it with a grain of salt.. there are some confidence bounds around those numbers but i dont have the calculates handy right now.
gbos, do you think one can accomplish measuring trendiness of underlying by using common tech analysis measurements instead of plugging OHLC data in XL much like what you did. I can think of at least the ADX indicator or maybe an RSI above 80 to indicate trend strength. What do you think this offers vs. canned indicators? Thanks
Nice XL sheet BTw.
Nice XL sheet BTw.
Of course one can use common technical analysis indicators to define trendiness. Depending on what he wants he can use different tools for different tasks. To give you an example Taleb was interested in trends for trading Barrier Options. So what was most important to him was the distribution of the extremes (the high and the low of the price during the day). So the Parkinson number tool made more sense to him than using a different indicator. Usually I think the simpler the tool the better unless you want to use something complex for specific reasons.
Found a very good paper here..
Extending the variance ratio test to visualize structure in data: an application to the S&P 100 Index.pdf
Extending the variance ratio test to visualize structure in data: an application to the S&P 100 Index.pdf
Why should that be a 'good' paper. Can't see how to make any money with this stuff.
It allows you to see trending vs. random vs. mean reverting series visually and provides pretty good tests for it. For insance, it is really useful for seeing if spreads are stable.