A paper -- Developing Robust Trading Systems

[howardbandy]

Greetings --

I have written a paper describing an original approach to evaluating the health of a trading system, and an original approach to determining position size based on system health that might be of interest to some of you.

It is entitled: Developing Robust Trading Systems, with Implications for Position Sizing and System Health

The technique described:
• Is of practical significance to practitioners of active investing.
• Produces both faster account growth and lower risk when compared with a passive buy-and-hold strategy.
• Is illustrated using a fully disclosed trading system, written in AmiBroker.

It can be found on the Resources page, in the Books category:
http://www.blueowlpress.com/WordPress/links/#books

Comments are welcome.

Best regards,
Howard

 

[captain black]

Thanks for posting that here Howard. Your willingness to share your knowledge and time both here and on the AB Yahoo group is much appreciated. Open discussion and a willingness to share and offer help is what keeps forums such as this important resources for Amibroker users and traders in general.

 

[Gringotts Bank]

Howard, on your website you link to a guy called Chris White who sells "EFT Bandit" system. Do you know him? Is he legit?

 

[tech/a]

Thanks for posting that here Howard. Your willingness to share your knowledge and time both here and on the AB Yahoo group is much appreciated. Open discussion and a willingness to share and offer help is what keeps forums such as this important resources for Amibroker users and traders in general.

While Im sure Howard is a nice guy
Dont lose sight that he has a vested interest (Business) in what he does.
We and I are customers.

 

[captain black]

While Im sure Howard is a nice guy
Dont lose sight that he has a vested interest (Business) in what he does.
We and I are customers.

Sure, but his willingness to openly answer questions and post code here and on the Yahoo group should be acknowledged. I think there's a difference between being helpful and blatant self promotion. Unlike a few others around I don't think Howard has crossed that line.

 

[tech/a]

Sure, but his willingness to openly answer questions and post code here and on the Yahoo group should be acknowledged. I think there's a difference between being helpful and blatant self promotion. Unlike a few others around I don't think Howard has crossed that line.

Yes I agree.

There is one thing I have noticed.
Howard doesnt get a great deal---if any flack on ASF
The reason I feel is he's a mathamatics PHD. So
the conversations seem to flow better.

 

[captain black]

There is one thing I have noticed.
Howard doesnt get a great deal---if any flack on ASF

When there is debate or discussion on technical issues like coding, system design etc then there really isn't any need for there to be any "flack" like there is on some of the more general discussion type threads where "personalities" come into play.

The Amibroker FAQ thread is a great example of somewhere where there is some great discussion and sharing of code and ideas. I've mentioned before that I have a background in open source software and have benefited greatly over several years from people generous with their time and knowledge. I think when someone (like yourself) is prepared to take the time to share their knowledge then it should be acknowledged.

There are a handful of regular contributors on the Amibroker Yahoo group such as Howard who have shared a lot of code that's now publically available. One of Amibroker's limitations is that the manual is to a certain degree not all that friendly for newbies. Being able to search online for solutions to coding problems often brings up answers in the Amibroker FAQ thread here, the Amibroker Yahoo group and also quite a few documents such as Howard's that help fill in the gaps.

I notice more and more now that a couple of the Indian based forums have quite an active Amibroker user base as well and a lot of searches are leading there. They like producing some very colourful charts but underneath it all there is some great work going on by some very talented individuals.

But at the end of the day it means nothing if were not making money, which I'd better get back to

 

[howardbandy]

Howard, on your website you link to a guy called Chris White who sells "EFT Bandit" system. Do you know him? Is he legit?

Greetings --

I have not met Chris White in person, but I have had many email and telephone conversations with him. Chris publishes two trading products -- ETF Bandit and Edge Rater. Edge Rater is more powerful, more general, and allows the trader / developer to do trading system design and testing. ETF Bandit is more focused. It implements the rules published by Larry Connors and Cesar Alvarez in their book "High Probability ETF Trading."

As you will have gathered from my speeches, postings, and books, I am an advocate of formula-based trading systems. Further, those systems must have trading characteristics such that they can be subjected to statistical validation. Mean reversion systems which trade frequently and have short holding periods, such as the Connors Alvarez systems, applied to liquid tradables fit this profile nicely.

Traders who want to investigate alternatives within the Connors Alvarez framework can read the rules in the book, program them into a platform (almost any platform is capable of handling the basic rules), and test various modifications. Traders who are satisfied with the rules as published and want to trade rather than develop can use the ETF Bandit program to give signals.

Based on everything I know about this (and I hope I will not need to issue a retraction in the future): The Connors Alvarez research published in the High Probability book is solid. The implementation of those rules in ETF Bandit is accurate. Chris White is a reputable businessman who has produced and is supporting a valuable and reasonably priced product in ETF Bandit.

I have enough confidence in what I have just described that I invited people on my mailing list to attend a free webinar hosted by Larry Connors in which he described some the research he has done. The webinar was quite informative, and feedback I received from people I know was positive. To be certain, Connors would like customers for his products, but there was essentially zero sales pressure during that webinar. I have enough confidence in Chris White's ETF Bandit that I plan to invite people on my mailing list to a free series of webinars describing how to use ETF Bandit to trade the Connors Alvarez systems.

I do have a business interest in all this -- I make a profit from books I sell and seminars I present. But my agenda is broader than profit. There is so much advice that is based on ignorance, on faulty logic, on subjective observation, or is intentionally misleading that I believe clearly stated alternatives based on sound mathematical and statistical foundations are needed. Everything I say is open and fully disclosed, so readers can replicate my results and determine for themselves whether the techniques I recommend are appropriate for their personal trading style. If that results in book sales, thank you.

Thanks for listening,
Howard

 

[tech/a]

I do have a business interest in all this -- I make a profit from books I sell and seminars I present. But my agenda is broader than profit. There is so much advice that is based on ignorance, on faulty logic, on subjective observation, or is intentionally misleading that I believe clearly stated alternatives based on sound mathematical and statistical foundations are needed. Everything I say is open and fully disclosed, so readers can replicate my results and determine for themselves whether the techniques I recommend are appropriate for their personal trading style. If that results in book sales, thank you.

No arguement here-----should be more of you!

 

[notting]

It's weird how people are always trying to come up with some new whiz bang sharpest edge trading system.
It's kind of old technology.
Has been done, can be tweaked fractionally here and there but basically, so much time reinventing what's invented for probably no more return.

It seems to be based on the premise that there is some secret system waiting to be discovered that will surely take advantage of all those gyrations.

I guess if your into that, great, play with your Lego, the point is lost on me.

Then again I find trying to pick the market entertaining!!!

 

[wayneL]

No arguement here-----should be more of you!

+1

Hence the difference in respect levels of someone genuine like Howard compared to others like Kertcher, Aussie Rob, and the like, who deservedly get worked over by us.

 

[tech/a]

It's weird how people are always trying to come up with some new whiz bang sharpest edge trading system.
It's kind of old technology.
Has been done, can be tweaked fractionally here and there but basically, so much time reinventing what's invented for probably no more return.

Well I've seen systems that make spectacular losses through to My best at 48%
to over 100%
Far from just a tweak or two.

It seems to be based on the premise that there is some secret system waiting to be discovered that will surely take advantage of all those gyrations.

You'll find that your ability to design trading methods that are profitable will put you way way ahead of the field who are discretionary trading.if you find trying to pick the market entertaining just imagine how much fun it will be when you can do it with regular monotony.

I guess if your into that, great, play with your Lego, the point is lost on me.

Then again I find trying to pick the market entertaining!!!

Take a second to run some figures
$100k at 15% compounded over 7 yrs
Then do
$100k at 50% compounded over 7 yrs
Then do
100k at 100% compounded over 7 yrs

Most traders make a loss.
If the figures you return don't have you questioning what your doing
Then Lego is certainly best for you.

 

[notting]

Yeah but I don't think there is a system that has ever returned better than 18 to 20% compounded on a rolling year basis over many years. That same system may have a 100% year on a leveraged account however if will give back and even out to around that 18 to 20 compounded over time.
Nothing wrong with it at all just never seen any beat it with proven consistency.
So don't see a point in trying to.
Unless its to make your own and sell it or something.

 

[tech/a]

Yeah but I don't think there is a system that has ever returned better than 18 to 20% compounded on a rolling year basis over many years. That same system may have a 100% year on a leveraged account however if will give back and even out to around that 18 to 20 compounded over time.
Nothing wrong with it at all just never seen any beat it with proven consistency.
So don't see a point in trying to.
Unless its to make your own and sell it or something.

Well there is
RADGE in his book Un Holy Grails shows many
His is 96% and it's not compounded and it's over many years.
I have one that's 48% Techtrader in the book is 28%

It's worth the effort

 

[notting]

Are you talking about the Radge Growth measured from 1 Jan 1997 - 6 Dec 2011 Averaging 30.99%?

 

[tech/a]

Pg 132
06-11
Only 5 yrs --- but.

 

[notting]

Better make the effort to download the book!!

 

[CanOz]

Yeah but I don't think there is a system that has ever returned better than 18 to 20% compounded on a rolling year basis over many years. That same system may have a 100% year on a leveraged account however if will give back and even out to around that 18 to 20 compounded over time.
Nothing wrong with it at all just never seen any beat it with proven consistency.
So don't see a point in trying to.
Unless its to make your own and sell it or something.

This utter nonsense. Spend an hour on Google and you can find plenty of trading systems that returned well over 50% annually. The turtles traded probably the most famous systematic approach.

Most successful henge funds are trading systematically. You don't seriously think there are traders that trade that well for that long using their instinct do you?

This is about using mathematics, its QUANTITATIVE, can you say QUANTITATIVE?

Here is the modified 20%Flipper on the Russell 3000. 2001 - 2012

CAGR anyone?

CanOz

 

[craft]

I can’t stand this without correction.

Howard’s paper discusses how to determine possible system robustness through Monte Carlo testing and out of sample simulation and then how to limit draw down in the face of a failing system in real life. Good work, seems a shame to derail his thread with ego driven crap.

Well I've seen systems that make spectacular losses through to My best at 48%
to over 100%
Far from just a tweak or two.



You'll find that your ability to design trading methods that are profitable will put you way way ahead of the field who are discretionary trading.if you find trying to pick the market entertaining just imagine how much fun it will be when you can do it with regular monotony.



Take a second to run some figures
$100k at 15% compounded over 7 yrs
Then do
$100k at 50% compounded over 7 yrs
Then do
100k at 100% compounded over 7 yrs

Most traders make a loss.
If the figures you return don't have you questioning what your doing
Then Lego is certainly best for you.

Show me a system that returns 100% that is not data mined to within an inch of its life, one that you would be prepared to invest your entire wealth in – and why wouldn’t you bet the house if it objectively returns 100% p.a.

100K at 100% means Just 20 years away from catching the best empirical evidence that discretionary value investing works.

Well there is
RADGE in his book Un Holy Grails shows many
His is 96% and it's not compounded and it's over many years.
I have one that's 48% Techtrader in the book is 28%

It's worth the effort

Radge's actual figures that Tech is quoting is 96.46% over a period of time. For an Annual Rate of Return of 13.12%. Tech Trader in the book is a simulated 21.33%
But why let facts get in the road of a good FIGJAM moment.

 

[notting]

Aware of most of that, turtles and all.