|
PureBytes Links
Trading Reference Links
|
Agreed.....great post.
In backtesting......regarding Max Drawdown.......watch-out for Oct 1997 and
Oct 1998 time-frames !!!
Your system may be terrific but the max drawdown may have occurred somewhere
in these times of high volatility and drastically fast down-moves.
Don't "kill" your system if it gets "hit" once in these "untypical" times.
Remember...Vic Niederhoffer was bankrupted in Oct 1997....in one day.....he
lost over 40 million.....shorting puts.
Note: I would say August 1987 is another back-testing "outlier"......very,
very few technical indicators could have seen that one coming.
> -----Original Message-----
> From: John Clayburg [mailto:clayburg@xxxxxxxxxx]
> Sent: Friday, November 19, 1999 9:47 AM
> To: fritz@xxxxxxxx
> Cc: omega-list@xxxxxxxxxx; Code List
> Subject: Re: CL_Evaluating & Optimizing systems
>
>
> Gary:
>
> Terrific post. The optimizer in TS can be the best or worst tool in the
> box, depending on how you use it.
>
> Your section on the dangers of optimization should be required reading in
> Trading 101.
>
> One thing I look at also is the number of consecutive losing trades. Seems
> like most people will abandon a system after 3 - 4 consecutive losses.
>
> I also mentally construct a net profit - drawdown ratio. Could do this is
> in Excel but haven't bothered.
>
> As a parameter, I like to see a system with at least 3x net profit to
> drawdown.
>
> John Clayburg
>
> Gary Fritz wrote:
>
> > Recently I've written up some "how to test systems" comments for a
> > few people. I decided this would probably be valuable for a lot of
> > folks, so I expanded it and (hopefully) made it more generally useful.
> >
> > If you are a system-development expert, please feel free to comment!
> > I think I have a pretty clear understanding of the optimization
> > process, but I'm very willing to learn more.
> >
> > Hope it's helpful,
> > Gary
> >
>
> <<snip>>
>
|