|
PureBytes Links
Trading Reference Links
|
New Version 7.0 eod?
Does anyone know if any of the system tester inadequacies will be resolved
in the upcoming 7.0 version? In particular my concern is with:
1) The inability to test using "a constant investment per trade" as you
mention Rudolf.
2) The very important option I find missing of being able to reference the
entry date and entry price from within the formula language. I do appreciate
Equis' small step in this direction of allowing different fields for entry
and exits as of version 6.52.
Best... John Rastutis
----- Original Message -----
From: rudolf stricker <lists@xxxxxxxxxxx>
To: <metastock@xxxxxxxxxxxxx>
Sent: Thursday, January 13, 2000 8:58 AM
Subject: Re: Rating trading systems
>
> Onno,
>
> thanks you for this interesting results.
>
> On Mon, 10 Jan 2000 20:10:51 +0100 (CET), you wrote:
>
> >The best system I have developed is the one for the Dutch stock index
(AEX).
> >I used period 1990-1998 for optimalisation.
> >
> >Period 1990-now:
> >Maximum profit: 5972 points (=100%)
> >My best trading system: 2118 points (35%)
> >
> >The rest of my trading systems score 5 to 15%.
>
> First: Do I assume right, that you used "reinvest 100%" in your tests,
> which is a Built-in MetaStock feature?
>
> If so: Imo this is not a very good measure for the profitability of a
> trading system, because _the sequence_ of the trading results has an
> important influence on the global result. This makes the results
> doubtfully in terms of generalization.
>
> Imo, a _constant investment per trade_ should be used in system test
> to avoid those "immanent & uncontrollable weighting" over time.
> Unfortunately, this is not implemented as a standard feature in
> MetaStock.
>
> Secondly: Did you trade the AEX directly?
>
> If so: Probably, this cannot be done in reality. So what about
> switching to options of the AEX, just to compare your results with my
> DAX results ...
>
> mfg rudolf stricker
>
> BTW: Did you do anything into the direction of generalization control?
> | Disclaimer: The views of this user are strictly his own.
>
|