[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[amibroker] Re: Your 5 most important backtest statistics



PureBytes Links

Trading Reference Links

--- In amibroker@xxxxxxxxxxxxxxx, "Michael.S.G." <OzFalconAB@xxx> wrote:
 
> I'm looking for the most common statistics used by people on their 
backtests 
> for the inclusion in the backtest framework. 
 
Could you explain a little more?
What is the backtest framework?
Has this been around before or is it entirely your own project?

> I was planning to upload the Backtest framework as some sort of 
> AmiBroker Community Open Source Project.

>From what I have seen proactive projects don't have a good track record 
in the forum but that might change.
I will help if I can.

My evaluation methodology is off the wall compared to others so it 
probably won't be of any use to you.

No offence to Ami, as it provides all industry standard evaluation 
measures, but I don't use them.
I only want the closed trade series and I export to Xcel for evaluation.
I have a custom built model in Xcel and I run it through that.
I evaluate everything differently to anyone I have seen, at least 
anyone who is prepared to comment about the subject in public.
The model is relatively new, and a work in progress, so it is partly 
manual at the moment.
Later I might pay a programmer to write it up as an Xcel plugin with 
some automation of the tasks.
I doubt if it is that unique as I believe the same evaluation can be 
done using MCS.
I prefer to use *my method* because I understand it from every which 
way and it *explains* evaluation to me in one single image.
Unlike MCS it is not processor hungry.

All trades are standardised to %.
The only recogniseable stat from Ami I use is win/loss ratio.
>From there I use the frequency distribution of the trades; but I'm not 
concerned about the distribution type.
I simulate the probable range of equity outcomes for different time 
periods and evaluate based on that.

I also consider significance, once again using my own method.

>From the above I arrive at a trade profile:

the probability of system ruin at time (t);
the probability that the system metrics obtained were due to chance;
and the probability of portfolio ruin at time (t).

If I don't like the profile I go back to the system and change the 
rules, which changes the key drivers and onto the profile.
MoneyManagement is inbuilt and the key drivers/profile feedback loop 
shows me if my MM is on the money.

Ruin is defined as 50% loss of capital, in which case I would walk away 
from the table with half of my capital intact.

The fun part is that I don't disagree with Hermans approach.
My image would be not unlike his rainbow curves.
I am also interested in the number of trades and that they come from a 
wide range of stocks.
Using intraday bars to up the trade count is another technique I have 
considered using.
It appeals, especially if the signal proves to be generic and stands up 
across time frames.

I have noticed before, that in trading, I can agree with two apparently 
contradictory methods or philosophies.
The trick is in understanding them enough to make them work and knowing 
which one you are actually using at the time.

I could hold my end up in the *stats can't be relied on* argument but I 
don't think I have the time right now.

I might have to have a go at carlacashes question if no one else does )
(god help us).


Brian_z




Please note that this group is for discussion between users only.

To get support from AmiBroker please send an e-mail directly to 
SUPPORT {at} amibroker.com

For NEW RELEASE ANNOUNCEMENTS and other news always check DEVLOG:
http://www.amibroker.com/devlog/

For other support material please check also:
http://www.amibroker.com/support.html
 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/amibroker/

<*> Your email settings:
    Individual Email | Traditional

<*> To change settings online go to:
    http://groups.yahoo.com/group/amibroker/join
    (Yahoo! ID required)

<*> To change settings via email:
    mailto:amibroker-digest@xxxxxxxxxxxxxxx 
    mailto:amibroker-fullfeatured@xxxxxxxxxxxxxxx

<*> To unsubscribe from this group, send an email to:
    amibroker-unsubscribe@xxxxxxxxxxxxxxx

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/