|
PureBytes Links
Trading Reference Links
|
David:
It's difficult to make conclusions based on 3 years of data on one stock.
Depending on the stock's characteristics and the time period, the exit's
parameters might need modification. I would suggest, though, expanding your
testing at this point rather than optimizing the parameters based on such a
small sample size.
Also, the Chandelier exit seems to perform better on systems that have a
little longer time horizon, so watch for a conflict between your entry
system and exit system.
The Chandelier exit isn't the holy grail, but the peace of mind of a solid
trailing stop (vs. no stop) can mean the difference between a tradeable
system (from an emotional point of view) and an untradeable one.
I'm not aware of a workaround for the PREV function. J suggested using
ref( ), but I don't think it would work in this case. The PREV function
certainly makes backtesting painful, though.
Regards.
----- Original Message -----
From: David Bozkurtian <dbozkurtian@xxxxxxxxxxx>
To: <metastock@xxxxxxxxxxxxx>
Sent: February 2, 2000 05:41
Subject: Re: Formula Question - Chandelier Exit
> Glen,
>
> I tested the C. Exit last night using Metastock. Did not seem to add
> anything to the bottom line. I was testing 3 years of data on a company
> called NEW SYSTEM. With the MS System Tester (Comparison mode) I came up
> with 4 systems which made decent money. I tried to change the exits on the
> most profitable system using the Chandelier Exit and got losses all
> around.
>
> I must be doing something wrong. I used the classic Equis MA penetration
> system. Not only that, but it took a really long time to calculate
> everything.
>
> Here's a side question. In VB and VBA, there are some functions in the
> language which should be avoided because they are too slow. There are
> workarounds to them. The PREV function in MS System Tester seems to be one
> of them. Does anyone know of a quicker (possibly long hand) workaround
> which will speed things up?
>
> Thanks in advance
>
> David
|