|
PureBytes Links
Trading Reference Links
|
Mark Johnson's analysis of the difference between the total profit for Fixed
Ratio and Spear "bet methods" is interesting, incomplete, and meaningless.
The difference in total profit between any two methods can never be
validated without knowing the variance of the data. Is the difference larger
or smaller than the standard deviation of the difference? Mr. Johnson
doesn't tell us.
Please perform the usual dependent t test to validate or invalidate the
results. Simply take the difference between the results of the two systems
for each contract ( 15 in total ?). Average the results. Compute the
standard deviation of the difference over the n = 15 contract sample. Divide
the average by the standard deviation to compute t. If t > 1.35, or < - 1.35
the systems differ with 90% confidence. If t > 1.771 or < -1.771 then the
systems differ with 95 % confidence.
Another problem that is not explained by reference to the code attached
to Johnson's message. How is the volatility computed in Spear's method? What
does " volatility less than 4% of the Equity" mean?
Inquiry the "Fixed Ratio" method uses a simple square root formula
compared with Vince's linear formula. Has anyone tried any other power of
Protit = Equity - Beginning Equity ? How about the Log ? And isn't all this
just a silly computer simulation diversion? Can you imagine "investing"
millions of dollars on coffee contracts? More likely any investor who makes
some money in commodities will re-invest in high flier's like T and US(30Y)
(cash not futures!).
Paul Lasky
|