Discussion in 'Data Sets and Feeds' started by Sergio77, Jul 8, 2015.
Interesting analysis: http://epchan.blogspot.gr/2015/07/time-series-analysis-and-data-gaps.html
This guy states:
"Physical laws may be unchanged since a few seconds after the Big Bang, but financial laws change every few years! I believe we should not use any data prior to 2009 for most analysis. The reason? Please look at Figure 5.10 of my second book Algorithmic Trading."
Do you have Figure 5.10?
Lastly, the comments indicate that maybe 2009 data is due to QE, which may therefore infer "ignore everything with QE" rather than "ignore everything before 2009"
I strongly suggest that everyone listen very closely to what Dr. Chan has to say. He knows his stuff! surf
LOL. His Kelly stuff is ridiculous.
Dr. Chan does not like the data before 2009, so the stock market (pretended) did not exist before 2009.
re: "but financial laws change every few years! I believe we should not use any data prior to 2009 for most analysis"
This was a ridiculous statement. Just because the markets change, does not preclude one to skip backtesting in prior periods.
Why ? Because the market could revert back to price action in that period of time !!
This is exactly why a lot of "systems" eventually fail: the market changes and the system has no logic to detect this change or the developer did not allow for a parameter to accomodate this change.
If you ignore QE you ignore one of the best uptrends.
If not ridiculus I would call it very risky. If he follows that he risks ruin.
Of course it's ridiculous. One of the advantages of true Kelly sizing is that there is never risk of ruin. But Chan has accepted without question Thorp's approximation, which is an approximation of an approximation. This compounding of errors leads to a completely nonoptimal and dangerous outcome.
The best Kelly formula is found here:
Separate names with a comma.