|
PureBytes Links
Trading Reference Links
|
<x-flowed>
I posted a message on the 20th for my Trading partner (Steve /
slwieringa@xxxxxxx) and forwarded the responses to him. He's since gone
over the responses, done some research and put together this followup
(attached below).
Thanks to everyone who took the time to respond.
Dave
his followup...
- - - - - -
From: "slwieringa @ AT&T" <slwieringa@xxxxxxx>
To: <David_wieringa@xxxxxxxxxxxxxxxx>
Subject: Summary: Simulating margin in your mechanical system
Date: Sat, 30 Sep 2000 12:57:42 -0400
What is this about?
This message summarizes my findings while searching for a method of
simulating the changing value of margin in my trading system.
Note: I have tried to lay this out clearly, but it should not be taken as
gospel - only my opinion. Many will have a different perspective and/or use
different terminology.
Who may be interested?
+ Anyone who would like to develop an 'estimated margin' indicator
+ Anyone who feels a quantified value of market volatility could improve
their trading decisions
+ Anyone who uses a purely mechanical trading approach
Why is using a simulated margin important?
A few reasons are:
+ If your trading rules use fixed dollar amounts the 'ideal' values will
vary dependent on market volatility
+ If your trading system is purely mechanical, it may need to adjust to
changes in volatility (using percentages offers some improvement here over
fixed dollar amounts...)
+ I'm not sure it's feasible to find an accurate and economical source of
historical margin data. Even if it exists, I rather not go down that path.
(One of the exchanges I checked with said they do not make this data
available on their web site. They did offer to answer any historical
margins questions if the commodity and dates were provided.)
What is margin, what does it represent, and how do the exchanges compute it?
Basically initial margin is the amount of available money you must have in
your account in order to buy or sell an additional contract. Maintenance
margin is the amount required to keep your contract. Margin is useful in
the client/brokerage relationship to protect the brokerage from losses. It
is my understanding that the exchanges set the minimum margin amount and
brokerages are free to adjust from this number for their own protection. In
contacting one of the exchanges I was told margin values are determined by:
1. Current market volatility
2. Historical market volatility
3. Their subjective predictions of future market volatility
What did I find?
+ Below I explain how we simulated maintenance margin. In comparing my
value to known historical values, I found that there can be a large
difference between using a Continuous Contract verses using a 'real
contract' particularly towards the end of the contact life. For example
during a market rally top of CBOT Corn during 1996, my simulated margin
using the real July (N) contract during the month of July was between $1260
and $1600. The exchange informed me the real historical values of margin
were between $900 and $5000 (no margin - cash to play!) Using a Continuous
Contract during the same time period the simulated margin was a bit lower
$880.
+ The following link will bring you to the section of the CBOT web site
dealing with margin. I am not sure if the CBOT is just stating what the
current minimum margin is, or if they have a new policy which will control
the minimum level? My guess is, I am just find the wording confusing and
the first is the case.
http://www.cbot.com/cbot/www/page/0,1398,13+49,00.html
Explanation of my code:
Unfortunately for all of you, my trading partner and I do all of our coding
in C++ and Java, so I am not able to easily share the code with you in a
format that would be useful to most of you. But our approach it is pretty
basic, so I will do my best to describe it.
First, from the exchange, they are interested in current volatility. To
represent this component, we calculate a value called Daily Volatility
which is equal to the greater of (end of day High minus previous day's
Close) or (end of day Low minus previous day's Close). Note: At first, I
compared end of day Close to end of day Close, but I think this works
better and is more accurate to the exchange's concerns...
Secondly, the exchange is interested in historical volatility. To represent
this component, we calculate a value called Historical Volatility by
looking back over a period of time (I am using 66 days/3 months) and find
the largest values of Daily Volatility. Since I don't want to put to much
emphasis on any one value (i.e reduce noise) I take an average of the (3)
highest values.
As you can image, this value of Historical Volatility is useful in itself
for reasons of data mining. But to convert this value to margin we just
need to multiply it by the commodity unit/dollar ratio. For example in CBOT
Corn, if the my historical volatility value is 9.0 cents and I multiply it
by $50/cent my simulated maintenance margin is $450.
I hope you find the this information useful to your trading.
Thanks again to each of you that contributed to my search for a solution.
david_wieringa@xxxxxxxx
Software Engineer (contractor)
Scottsdale, AZ / Grand Rapids, MI
</x-flowed>
|