|
PureBytes Links
Trading Reference Links
|
My experience has been quite different, it seems.
The difficulty we encountered in estimating max DD was in the report it's
self. I'll give you an example. We ran a series of optimizations on TS6 on
three different machines. (two with exactly the same hardware) All the
cache data was flushed and downloaded for the @ES.D. Despite this, all
three machines had very different reports and vastly different equity curves.
Recently, we attempted to verify tests done several months ago, again on
the @ES.D. The strategy went from a profit to a loss using the same date
range and the same settings on the strategy.
My understanding, from Sam Tennis and techs at Tradestation, is that the
report by Rina Systems has had some unknown flaws for years. So, given the
lack of consistency in the reporting, how would you estimate max DD?
Not to mention, that the reporting is TS's SOLE reason for existence! If it
has flaws, or can't be supported by TS (earlier than the recent Beta
version offering) WHAT IS THE USE OF THE THING???
At 12:10 AM 4/30/03 +0200, you wrote:
In my opinion, the right way to measure DD is the way Daniel described it
(or Tomas Stridsman) but I dont think the question about "what to do ..."
if the DD breaks the measured Max DD have been answered totally.
I had planned to bring this question up on the list because I have thought
about it so it would be interesting to hear from more people´s.
Regards
_______________________________________________
Hi Todd,
in my opinion is the way you describe the Drawdown totally wrong.
You shouldn´t give the Strategy Report in TS any value because it just tells
you the $ amount of the drawdown- not when and where ist happend....
You should learn to think in terms of % rather than in terms of $.
Let me give you an example:
Lets say you bought a stock at 100$. You sell it at 90$. So your drawdown is
after this trade 10$ or 10%.
The next time you buy that stock is at 200$. This time you sell is at 190$.
So you drawdown is again 10$. But the main and most important difference:
this drawdown was only 5%- half the drawdown when the stock traded at 100$.
So, when you test a strategy the $ amount of drawdown is useless beause it
doesn´t tell you when and where ist happend.
Best wishes from Bayreuth,
Daniel
----- Original Message -----
From: "Todd Hoff" <gatrboots@xxxxxxxxx>
To: "Omega forum posting" <omega-list@xxxxxxxxxx>
Sent: Monday, April 28, 2003 12:44 PM
Subject: when a strategy breaks the max drawdown
> Hello Everyone,
>
> I could really use everyone's thoughts on this topic
> which always seems to happen to me. Sorry if this
> topic has already been discussed:
>
> In your opinion, when or by how much do you
> realistically stop trading your strategy when it
> breaks the max historical drawdown?
>
> For example. Let's say, that I back-tested my strategy
> for the last 5 years and the max historical drawdown
> on this e mini sp system on a 1 contract basis was
> $3,000. I was once told that we should give it some
> leeway when ,in real time trading, it approaches the
> max drawdown. Someone told me once to give it 20%. So
> in this example, we would stop trading the strategy
> when it reached $3,600. Others have told me give it
> 50%.
>
> I don't know if it is just murphey's law or just some
> not so great systems, but every system that I have
> traded, always breaks the max drawdown substantially
> and it usually falls apart after trading it real time.
> It seems that we just never know how much to allow
> the system to exceed the max drawdown.
>
> Sometimes I have stopped trading a system too early ,
> when it barely broke the max dd and then recovered and
> other times, I might have given it even too much
> leeway when it broke the max dd, since it continued
> drawing down even after stopping it.
>
> Your input will be very welcome and helpful to me.
>
> Thanks,
>
> Todd Hoff
>
|