|
PureBytes Links
Trading Reference Links
|
holy grail for me is:
- 60% or higher
- 1 to 1 or higher...
- no parameters or only one parameter = time frame ( fully adaptive )
- works on any time series
the rest of the stats will come from the 60% and 1 to 1...
cuz if you have a system with 60-1-1 means drawdown is limited,
eq is rising steadily with ok volat.... basically the rest will fall
in place... if those two stats are good.
your graph is same graph i posted sometime ago on system only, ie
the breakeven curve concept...
only trade systems that fall above the curve, above 1 and
to the right of 60%...
and i can only see one dot there :-) what's the name of that one?
bilo.
----- Original Message -----
From: "Ivo Karindi" <ivo@xxxxxxxxx>
To: "OmegaList (E-mail)" <omega-list@xxxxxxxxxx>
Sent: Monday, June 11, 2001 2:56 PM
Subject: System Performance (philosophy)
> My understanding about effectiveness of the statistics used to describe
> trading systems' performance boils down to 2 numbers: average win/loss
ratio
> and % of winners. These are the numbers that are capable of describing a
> system's performance over various timeframes and markets in a relatively
> constant manner. The rest of the numbers like max drawdown, net
profit/loss
> etc. are rather probabilistic and limited in nature, meaning that they
> either apply to one single timeframe/market, or to some system-adverse
time
> periods.
>
> Based on this, I took the numbers from the recent article in Futures
> Magazine about top 10 trading systems (thanks Mr. Simms), put them into an
> Excel spreadsheet, and charted all the systems based on the above 2
numbers
> (attached, the green line is the breakeven borderline). Talking about the
> "Holy Grail" trading systems, most of the systems based on these numbers
> clearly do not qualify into this category. Moreover, even the systems
that
> look really good obviously performed so well only due to some very nice
> bull-markets.
>
> Now, do these "Holy Grail" systems actually exist in reality? And what is
a
> "Holy Grail" system? 70% winners with avg win/loss ratio of 3-4? Is this
> possible as a rule for a system, not just a temporary exception? (Anyone
> wants to share their "Holy Grail" experience?) Or is the whole game in
> having a slight edge over completely random results, and then applying a
> sound position sizing / risk management strategy?
>
> Ivo
>
>
>
>
> PS If anyone wants the spreadsheet just let me know.
>
|