"edge"

When someone would come to me

with his track record, I wouldn't be

interested in his return only in the

significancy of his trading......

I had a discussion with my prof of

statisctics, he tols me that if my signi-

ficance level would be at 0,1 % , he

would be interested in letting his

money mange by me.

I don't know if it is that good , I still

have to calculate it.......

He wasn't asking for my return.....

very smart guy, that guy....
 
Once you have calculated the significancy

you can starting to plan the degree

of leverage to use.


The smaller the significancy, the more

aggressive you can be in your money

management plan.

With Monte Carlo simulations you

can then determine confidence levels

for specific drawdown's to happen.

Another important aspect is how

stable your edge is over time.
 
I think the discussion itself shows that everybody has a different definition of "edge". I think that edge itself is a very personal concept. If I become better than what was an edge to me yesterday might not be one today.

edge means something really special, really true, not fading away. it increases personal strength to have an edge. loosing an edge is a horrible thing.

I think the universe of "edges" (first time I use the plural) is highly inhomogeneous. A thing can be more or less "edgy". I think the root for quarrels on this is that people with more edgy tools laugh about those with less edgy stuff, which makes the latter angry and so forth.

A stupid "trading system" I did on the SP on daily data only shows the following results since 1997. It is long only and horribly simple.

1997 33.5%
1998 3.0%
1999 7.5%
2000 19.7%
2001 8.5%
2002 54.1%
2003 8.9%

it made 80 trades with a hit ratio of 69%. Wouldn't trade it but it is funny, since it was a by product of research on the edge thing.


peace
 
Quote from acrary:

Here's the basic code in case anyone wants to convert it to their platform of choice.
More...

I looked through your code and I dont think the calculation is correct for intraday systems. In particular, the following line is problematic:
prof = cl(ntrd+st+ho) - cl(ntrd+st)

Here you are calculating the "random day" profit/loss from close to close. However, if you are trading an intraday breakout system, where you buy at a certain price level then sell at close, the "random day" profit/loss should be calculated by randomly selecting a point in time during the trading day and then looking at the price difference from that point to the close.

In other words, I dont think that a random close-to-close calculation is an accurate baseline comparison for a intraday-to-close system.

-bbc
 
UNLESS your win rate is only 1%.
Then the equity curve is SCARY.

A combo of win% and PF = nice equity curve.

peace

axeman


Quote from traderkay:

ProfitFactor=(%Wins*AvgWin) / (%Losses*AvgLoss)
It should also incorporate commissions.
If PF>1, technically you have an edge. The bigger it is, the bigger your edge and the smoother your equity curve will be. For example PF of 2 will always give you an equity curve to kill for.
More...
 
Quote from sabena:

How would you quantify an "edge" ?
More...


Aside from inside or special information or special market privileges, I don't think that you can reliably quantify an "edge" with any meaningful accuracy. The people who think that they can do not fully understand the question or the vagaries of the market which is not accommodating enough to provide us with a meaningful probability distribution. The true probability distribution is neither known nor stationary. (Picture someone who takes two seat-of-the-pants estimates, multiplies them, and then gets accuracy to the 8th decimal point.) Therefore, we will have to do with vague, ballpark approximations at best.

As for what an "edge" is, it is easier to define what it is not. Most people who enter the markets think they have an edge. Clearly, they do not.
 
You must log in or register to reply here.
Back
Top