|
PureBytes Links
Trading Reference Links
|
Adrian
There is a very narrow defined band in the price movements where I can,
often to the tick, in advance tell where the market will turn. Took me
"only" seven years of full time study to figure it all out.
Once when a person knows this then it becomes a matter of: "are we
trending?", "are we in a chop?", "what is the market 'manipulator' <wink> up
to now?" (and not only the MM but also governments can interfere, very well
visible if you know how to look for it) and how many contracts do I trade?
The last question (of how many contracts) has a lot to do with position
sizing. If you know that the trend continues and that this present movement
is a retrace then you can add. In today's robot's traded environment (means
reverting) on short timeframes this is a very risky proposition that I am
not prepared to participate with. Lets look at the 28th Feb dip in the
market wiping out several months of gains. Even last week I could see
intervention.
The first question determine what trade I put on, if any.
I trade both sides of the market, let's not forget 1929 where the market
took, a what?, 90% dip? Could you have survived if you had been trading on
margin? And let's not talk about "it won't happen again". Got some strong
views on this.
But if position sizing works for you then good luck to you and well done.
Then you may be a better trader than I am.
I do not want to go deeper into this, it would reveal too much of what I am
doing in order to explain to you what is possible. Let's close this
discussion, apologies if I have offended anyone.
best
Marinus
----- Original Message -----
From: "Adrian Pitt" <apitt@xxxxxxxxxxxxx>
To: "'Loewensteijn'" <Loewensteijn@xxxxxxxxxxxx>; <omega-list@xxxxxxxxxx>
Sent: Tuesday 31 July 2007 15:40
Subject: RE: position sizing
Marinus,
1. Position sizing is not about where you enter, or in how many positions.
It is a portfolio risk measurement, where the size of your position is
determined by targeting either annualised return or risk to your portfolio
over the life of the backtest.
2. Yes, there is an optimal level to enter a position. It is 1 tick
above
the low on a buy and 1 tick below the high on a sell. Unfortunately, none
of us know where that is in advance, and even if you designed a model that
could enter at those points, it is no guarantee it will work into the
future. In fact you can pretty much guarantee that the market will change
its characteristics enough to result in your model deteriorating somewhat.
Therefore, almost without exception, every successful trader of large
magnitude will enter at multiple points. This is not only a volume issue
(which isn't relevant to a small trader), but also recognition of the fact
that the future won't repeat the past identically, and if the optimal
entry
point shifts back and forth over time, then by entering via multiple
positions, it generally helps keep you somewhat within the vicinity of the
optimal point. Entering via multiple positions has absolutely no
relevance
here to multiple models or averaging into losses, nor taking enormous
risks.
This is because the trader/investor sees the entries as one trade, where
risk is defined beforehand and determined to be within portfolio
constraints
on the total position.
Adrian
|