Which benchmark will it beat on a risk-adjusted basis?
Random Walk Theory Proved, once and for all.
- Thread starter mu200411
- Start date
Which benchmark will it beat on a risk-adjusted basis?
if you can develop a system to profitably trade the random data, then in all probability it is not random to start with. It is like using mathematical formulae to generate pseudo-random numbers (which pass all the usual tests of randomness), and proceeding to predict the next items in the sequence by your knowledge of the formulae.
no i said i could take random walk theory which is a formula a math formula and make it be profitable on any data find some random data or fabricate some make it as tricky as you wish you could make it cycle like a fft and then the new data could go strait up or down like a trend and i could capture it no matter..
riskaverse305
Guest
for a discussion of "random walk," this thread's lack of mathematics is astounding.
an interesting theorem for symmetric random walks:
if W(t) is brownian motion (represents the gain(loss) of a stock process at time t), and we define tau = time that W(t) hits price a or b (a>=b)
then
P[tau less than infinity]=1
i.e
take ANY two price levels around a starting price-- could be +/-1,000,000 -- and the probability that a SYMMETRIC (no trend) br. motion will hit one of these price levels is 1.
ito processes have there place though, even if br. motion is obviously inappropriate.
an interesting theorem for symmetric random walks:
if W(t) is brownian motion (represents the gain(loss) of a stock process at time t), and we define tau = time that W(t) hits price a or b (a>=b)
then
P[tau less than infinity]=1
i.e
take ANY two price levels around a starting price-- could be +/-1,000,000 -- and the probability that a SYMMETRIC (no trend) br. motion will hit one of these price levels is 1.
ito processes have there place though, even if br. motion is obviously inappropriate.
riskaverse305
Guest
Quote from MarkBrown:
i have mine show me yours
arithmetic?
