|
PureBytes Links
Trading Reference Links
|
Dear James (or is it Randy?):
Please forgive me for taking so long in replying. I have been just
plain swamped.
I'll try to be more specific. First, I have indicated to a few
respondents that I will describe my simulations and publish the
results on the Omega forum sometime in the near future...I'm committed
to it now, so it's only a matter of time. I don't know that it's
anything earthshaking, but I figure I ought to since I brought it up
and it might be a good idea to submit it for "peer review." It's
probably nothing new and might even be a real letdown to some. Until
then...
Money management strategies generally base their allocations on the
amount of equity in the account being traded. If you double it in a
week, you have twice as much capital ("available capital") to
allocate. For example, if you are doing 1 DMark for every 10,000 in
your account, then the number of DMarks you trade will vary in direct
proportion to the size of the account. Whatever schemes you come up
with to vary the number of contracts, in general, the final number
varies directly with the account equity ("available cap. = account
equity). I assume that this is the way most money management
strategies begin.
This is perfect common sense and there is nothing wrong with it.
Where this might get scary is when your equity is hitting highs and
your position size/exposure increases directly with it. Then
monumental drawdowns are possible.
OK, so my variant of Optimal f (which ought to work with any
fractional reinvestment approach) also varies the "available capital"
with the overall account equity, but not directly. It's some function
of it (i.e. Available equity = f(account equity)) that would require
more of a math guy than I (which doesn't say much) to derive. I only
know what the results looked like for my simulation.
Instead of basing allocations on the actual trading account equity, I
based them on a "parallel account" which "trades" (on paper only)
single contracts only. This is the dummy account, but I like the term
"parrallel account" or even "Virtual account" better! So, the
"available equity" is the single account, and as it varies, so varies
the equity available for compounding. I hope I'm making it clear
enough without leaving anything out. Instead of compounding total
profits, you compound a part of them.
In my simulations, available equity was reevaluated daily and
positions adjusted accordingly. If one were to do this with the full
account, then a good run would produce a parabolic increase in equity
with an inevitable cliff dive.
By restricting the amount of capital availabe to compound with, a more
conservative equity curve (and TWR) is achieved.
My simulations with simply varying the percentage of the f factor did
not eliminate the parabolic/cliff dive sequence. Buffering the
"available equity" did. That's why I got excited about it.
NOTES: I have only tested this in simulations. The system was
nothing special. If you'll bear with me, I'll try to write up a
comprehensive explanation complete with pictures and stuff.
Please let me know if I've answered your question. Also, I'm going to
post this to the Omega list. Listers, if you've read this far, does
it sound useful enough to justify my posting the simulations?
Thanks for your interest.
Sincerely,
Charles
---James Rehler <jrehler@xxxxxxxxx> wrote:
>
> >From Omega-list
>
> You wrote:
>
> *******
> During one test, I accidently created a dummy single contract account
> which was referenced for capital availability, and the compounding was
> based on it. The result was a radically smoother equity curve than
> any of the straight Op f tests, but of course a radically reduced TWR.
> Still, I am excited about its potential. It would take someone
> smarter than I to describe exactly what I did mathematically, but it
> is tradeable (that is, it can be done in real time). As I am not
> trading currently I am not using it. But I look forward to getting
> "out of the lab" and putting it to use.
> *******
>
> Could you be more specific about what you mean by referencing
available
> capital?
>
> Randy
>
>
|