System Development with acrary

[acrary]

Quote from manz66:


So, do we need a system of more than 50% winning? I thought even money system can have positive expectancy using money management rule. Or do we need edge in a trading system? And, by how many percentage?

Yes, the risk of ruin is alive and well.
I tried to show that the win % is not very important toward the goal of consistent profitability.
Yes, you do need a edge to avoid risk of ruin. Otherwise as you increase the number of trades in random trading it's just a matter of time before you hit a drawdown from which you'll never recover.
Money management cannot improve the profitability or consistency of a method. It can multiply the effects of whatever edge the method has so that you can compound the profits at a rapid rate. There are however some position management techniques that can improve a marginal trading method. I'll do a series of posts on them at a later time.

 

[mind]

my thinking i terms of system is illustrated best by an example:

http://www.elitetrader.com/vb/showthread.php?s=&threadid=31847&highlight=hang+seng

still do not know what to think about this.



in terms of trading frequency i think you are perfectly right, alan. if you can shrink the times between trades while keeping the positive statistics alive (obviously against their natural enemy: trading costs), you end up with better stats. just take a mod SR of 1 for a normal trend follower over twenty years and assume you can have this shape of the curve every day. followed by another day with the same shape. makes drawdowns vapour away.
yet the challenge is keeping stats alive while trading high freq. you know olsen? trade SR 3.9, one of the highest numbers i have ever seen for a quant fund.

peace

 

[mind]

i agree on the edge test. i definitely did something very different. thanks for being that clear.

peace

 

[mind]

...

 

[acrary]

Quote from mind:

...

I think I understand where you're coming from. If I'm right it appears your interests are in the mining of the data to find opportunities. If this is your interest you should have this book:

http://www.amazon.com/exec/obidos/tg/detail/-/0691042896/103-0088184-2213474?v=glance

It's the bible of Time Series Analysis.

I think of Sharpe Ratio as a measure of risk. The lower the number, the greater the risk. High frequency trading trys to improve the SR by increasing the frequency. I add to that the idea of improving the profit factor as a way to improve overall results. Obviously you could create a SR above 4 just by using targets (profit versus loss), and by creating complimentary strategies (pairs trading, using stop loss of one method as the setup for entry in the opposite direction using another, etc.). I believe the best balance is to improve both frequency and quality of each trade to the largest extent possible. Then mix non-correlated methods together to improve the SR. After that use MMgt tools to manage the risk (which I define as max dd).

Hopefully we're not talking past one another. I find your work interesting even though it requires a paradigm shift in my thinking to try and understand it.

 

[mind]

i have the book i think.
i profit from this communication a lot. i am the disciple. and i am trying to do my homework.
i agree very much in all you concepts on high frequency. consider your equity curve tested back over ten years happening within one day. and tomorrow and ...
that illustrates the difference that high frequency makes.

peace

 

[damir00]

Quote from mind:

always assuming worst case means that you do not trade at all.

i don't see how that follows. there are a million financial instruments out there, knowing what the worst case looks like allows you to choose exactly how much - and what kind of - pain you're willing to endure. look at it another way - your statement implies blowups are inevitable.

 

[bdixon619]

Quote from acrary:

I read his interview in the Oct. 97 issue. Doesn't seem to offer any value from what I see: http://www.traders.com/Documentatio...97/Abstracts_new/Interview/Interview9710.html

I ordered his book that was published in 1997. I'll check it out when I receive it. I have the book he wrote with Kroll and I didn't find any value in it.

In an earlier post you said you learned and forgot about this material. I was just looking for some insight into what you replaced it with. I've been using it since 1988 and haven't found anything better so I was deeply interested.

Wow, impressive initiative! I don't know if the excerpt gives much info relative to what you posted. I have that issue, naturally. I will make a post with some info from the article as it relates to your post.

It is just possible I am on a tack unexplored by you, that of using some ideas from quality control. Some of those ideas influenced me when I was preparing the info I shared in my Journal.

After I read the Chande article and while I was still trading stocks (equities) I used some of the things I mentioned in my post exchange with Student to get a better idea of the efficient uses of limited capital. I will make a separate post about that.

I have to be away several hours today and am doing other things as well so I will probably get just one post done by the end of the evening.

 

[mind]

Quote from damir00:

... your statement implies blowups are inevitable.

funny you say that. we had a working hypothesis that no matter how good your strategy might be, as soon as it is not riskfree it will soner or later face this most unlikely streak of events that make it go bankrupt. having said that i must state that this could be in five billion years for one startegy and ten thousand for another. but the basic concpet was: measure system quality as the expected time until ruin. the longer the better.
i must add that within my group we actually did not completely agree upon the base assumption that eventually everyone will go bankrupt.

nevertheless after having skipped the psychological negativity of the subject, it might me a reasonable concept to describe risk.

 

[opmtrader]

Acrary, thank you as always for contributing to our education as traders. You are a beacon of light in the fog.

I noticed that correlations are used in much of your work. I was wondering if you took sample size into account when performing these correlation calculations. I now use a "correlation rank" in my work after reading the following in Victor Niederhoffer's Practical Speculation:

"In practice, we have found any number above 0.10 or below -0.10 based on 100 or more observations to be useful. The general formula we follow for usefulness is that the correlation coefficient times the number should be greater than 10. Thus, for 50 observation, a correlation of 0.20 would be useful and for 20 observations, a correlation of 0.50 would be useful." (pg. 196 to 197)

Lastly I would like to commend you on what I feel is one of the most pivotal and most overlooked aspects of systematic trading; organization. Being organized and systematic in your testing procedure, portfolio management, and trading implementation seems to be the key to success and longevity. I am working towards being more organized in my approach. Thanks for showing me what is possible.