|
PureBytes Links
Trading Reference Links
|
:I read it, and generally understood his point that 'modern portfolio
theory'
:is lacking, but while telling the reader that he believes that it is indeed
:possible to anticipate huge spikes in profit or loss using his methods, he
:doesn't seem to say _how_ to do this. I assume he would suggest using some
:sort of pattern recognition software to identify a 'generator' and then use
:his math to figure out possible future moves (multifractals) based on this?
:I'm not sure if I got this right, so if anyone else has a clearer picture
of
:what he was getting at, I'd love to hear it, cuz I was left pretty confused
:as to how to apply what he talks about.
My impression was that he says that the price movement / range/ depth of a
move was...."pre-determined?" (my attempt at an interpretation) but the time
frame in which it moves is variable. That was the impression I got from
seeing his 2nd graph and his explaination of it.
As for how to use it...I think that Clyde Lee's Swing Machine may be
applicable.
We can probrably use Fib ratios to make price projection on how large a
price movement might be but then use some statistics on Clyde's LOS to
determine the time frame. For example, a Fib ratio of 100 points may be
projected but the time it takes to make that move is indeterminable (it may
take 10 bars or 100 bars). Using a distribution set on Clyde's LOS, we can
say that in the past, a swing of 100 points took on average 5 bars. So it is
likely that the next swing will take 5 bars or so. Or we can say that this
last swing took 5 bars and moved 100 points. It is likely that this swing is
over and a new one may begin.
Or something like that. If anyone can determine a way of using his
article/information, please keep us inform.
|