|
PureBytes Links
Trading Reference Links
|
I have a book in my library that's called "Value Averaging" by Michael
Edleson, I think he's a professor of statistics also. The book was written
in 1991, and the basic concept is that one should add an equal amount of
value to your portfolio on a fixed basis.
1st month, add say ...10000
portfolio value 10000
2nd month, add whatever is needed to bring the value of the portfolio to
20000
3rd month , " " " " " "
" 30000
and so on...
The method buys more shares when the price is down and fewer when the price
is up. He even suggested that if actual account value exceeded the target
value for that period, that you should sell enough shares to bring it down
to the target value forceing you to sell at what is probably a high point in
the cycle.
The authors purpose was to present a method that was better than dollar cost
averaging and goes to length proving it.
I'm also always looking for new methods I can use. I appied this method with
a daily instead of monthly time period and invested 1/8 th of my paper
capitol in just one stock with a good future (fundamentaly strong and timely
according to value line) that had a high beta. When fully invested ( at
between 3 to 10 days) I put a tight trailing stop on the whole one stock
portfolio at 1/4 point below the previous days close and adjusted it daily
till being stopped out. I would then repeat the cycle.
The transaction costs are high, but the mechanical approach takes a lot of
the emotion out of it, I've never used it in real life but I thought it
sounded similar to what you were describing and the book might be some help
in what your developing.
Tom
----- Original Message -----
From: <TWA7663@xxxxxxx>
To: RealTraders Discussion Group <realtraders@xxxxxxxxxxxxxx>
Sent: Monday, May 24, 1999 4:11 PM
Subject: Money Management
>I have heard from several traders on this list or other trader lists that
bet
>size based upon a martingale method was stupid unless it was a reverse
>martingale (bet more on wins, less on losses). BTW, my use of the word
>"martingale" merely means that bets are varied by an amount that is
>determined from the previous wins and losses. It can be any algorithm, not
>just the stupid "doubling down" after a loss. I also define a regular
>martingale as a method that requires buying more after "some, not
necessarily
>all" losses; whereas, a reverse martingale often requires buying less after
>"some, not necessarily all" losses.
>
>About four years ago, I spent a couple of hours on the phone with a retired
>statistics professor that had also spent a lot of time working with some
>missile company. Please excuse that I can't remember his name or all the
>facts about the professor; however, I "do" remember he .....
>
>1. Traded a significant stock portfolio.
>2. Traded only long positions
>3. Used a sophisticated modified martingale. It was a unique algorithm
that
>often purchased more shares when the price went against him. It was always
>in the market unless certain predetermined circumstances occurred (rarely
out
>of the market because of the stocks chosen to trade). I could explain it
in
>detail but it would take a very lengthy discussion. We had fun talking
>because I had been kicked out of the Tahoe casinos when I was using my own
>modified martingale. They thought I was counting cards.
>4. Claimed to have had huge annual returns for several years. If I
remember
>correctly, over 50% annual on some stocks that had gone sideways or even
>lower for a year.
>5. Chose stocks that had high short term volatility relative to their very
>low longer-term volatility.
>6. Said the concept would work with any market and that he had a friend
that
>was making multiples of his returns with OEX options. He chose low
>volatility stocks because his risk was extremely low.
>
>I told him about the negative remarks about martingale systems that I had
>heard from other system gurus. He said that most just did not know how to
>minimize the risk incurred from martingale systems and did not know how to
>enhance the positive features. His algorithm went to great lengths to do
>this.
>
>I became quite excited with our discussions and spent a lot of time with
>Excel to test his concepts as well as other martingales. I even hired a VB
>programmer to help. I abandoned the project because others on this type of
>list and friends discouraged me. In addition, it was going to cost
thousands
>more to develop code for what I considered an adequate method to test the
>concept. I feel most comfortable trading systems that I have tested that
>generate 100000+ trades. I realize that may be overkill, but that gives me
a
>psychological edge. Since I was still very "green" at trading and easily
>impressed by negative comments, I abandoned the idea. However, my very
>preliminary testing seemed to show that the professor's method and some of
my
>own variations worked great. Many times I wish that I had continued.
>
>Has anyone tried to develop martingale ( that may buy more after a loss)
code
>that will simulate thousands of trades across multiple markets?
>
>Has anyone actually traded this way?
>
>If someone can answer yes, I think their experience would be of great
>interest to many of us. In addition, if there is a proficient VB
programmer
>out there that wants to help develop a system, I can share what I learned.
>I thought about developing EL code with a "C" program linked to it, as I
have
>done many times. I have spent thousands on a "C" programmer that is also
an
>EL expert. However, I think it is just easier to use Excel because of some
>of the crazy problems encountered with Easy Language. I have a problem
>trusting EL results.
>
>Like most on this list, I have heard many negative reasons why NOT to use a
>system that may buy more after a loss. I hope we could benefit by limiting
>the discussion to those that have tested or traded with success by buying
>more after losses. If we get no further discussion, we can assume that
none
>has or none knows of anyone that has had success with this method.
>
>Russ
>
>
>
>
|