Trading System - Macro vs. Micro

[brettc4]

I am hoping for some words of wisdom regarding Macro Vs Micro Trading Systems.

People are searching for the one system to rule them all (shocking Lord of the Rings reference there). People talk about back testing, and they talk about the number of securities it is tested against.

You also hear people talking about sentiment and psychology of the crowd.

Each security shares some characterictics with all others.
Each security in a specific sector shares additional characteristics with other shares in it's sector.
But then each shares earns income in slightly different ways even in the same sector.
The share holdings are different.
The size of the company will dictate which funds look at the stock.
Dividend yeild will dictate which funds look at the stock.
The type of the majority share holders ie founders v's funds will affect the liquidity.

** Here is where I am looking for wisom and guidance **

Why should I not select a small group of stocks (but large enough that entries are triggered across the group regularly) and then for each of these stocks look at the entry and exit conditions that work best for it, regardless of that being overbought/oversold condition, MA crossovers, Trend breakouts support/resistance whatever.
Backtesting this particular entry and exit against this stock and then following the plan with revision when necessary.

Thanks Brett

PS if this has been convered before could someone direct me to the post because I can't seem to find it.

 

[MS+Tradesim]

Re: Trading System - Macro Vs Micro

Brett,

Doubtless it has been covered before. And you can build a system which only trades a few stocks (whatever your definition of "few") and is highly optimised for just that basket. But the fewer the securities the less signals that it will generate. This may not be an issue for you depending on your objectives (something you need to ascertion prior to designing a system). You will also need to be right on top of the system to modify it as the securities change character and the system breaks down (the major problem of highly opimised systems). But a high expectancy system with few opportunities will probably not be as profitable as a lower expectancy system with more opportunities.

For example consider a very long-term trend following system which has an expectancy of 2R and generates 10 trades in a year. Your return will be year 20R. But how about a system of 0.4R which generates 120 trades a month? Your expected return could be 48R.

I have two systems, one medium term and one short term and the only price filter I use is > $0.02 so they generate far more trades than I can take which means my opportunity is limited only by capital.

 

[MS+Tradesim]

Re: Trading System - Macro vs Micro

Oops,

"But how about a system of 0.4R which generates 120 trades a month?"

should be... "But how about a system of 0.4R which generates 120 trades a year?".

Really don't like the lack of edit capability.

 

[nizar]

Re: Trading System - Macro Vs Micro

Brett,

Doubtless it has been covered before. And you can build a system which only trades a few stocks (whatever your definition of "few") and is highly optimised for just that basket. But the fewer the securities the less signals that it will generate. This may not be an issue for you depending on your objectives (something you need to ascertion prior to designing a system). You will also need to be right on top of the system to modify it as the securities change character and the system breaks down (the major problem of highly opimised systems). But a high expectancy system with few opportunities will probably not be as profitable as a lower expectancy system with more opportunities.
For example consider a very long-term trend following system which has an expectancy of 2R and generates 10 trades in a year. Your return will be year 20R. But how about a system of 0.4R which generates 120 trades a month? Your expected return could be 48R.

I have two systems, one medium term and one short term and the only price filter I use is > $0.02 so they generate far more trades than I can take which means my opportunity is limited only by capital.

Solid post MS+TS.
Worth repeating.

 

[Joe Blow]

Re: Trading System - Macro vs Micro

Really don't like the lack of edit capability.

Posts are editable for up to 20 minutes after they are submitted. Please take advantage of this time to proofread your post and make any necessary corrections.

 

[new girl]

Re: Trading System - Macro vs Micro

Posts are editable for up to 20 minutes after they are submitted. Please take advantage of this time to proofread your post and make any necessary corrections.

I realise that there aren't many women around here But being the woman that I am I keep changing my mind. 20 minutes is simply not long enough, is it possible to make it 30. please pretty please

And also I love using smilies and 5 is also simply not enough how about ten at least for me if the others don't care

 

[Joe Blow]

Re: Trading System - Macro vs Micro

I realise that there aren't many women around here But being the woman that I am I keep changing my mind. 20 minutes is simply not long enough, is it possible to make it 30. please pretty please

And also I love using smilies and 5 is also simply not enough how about ten at least for me if the others don't care

Don't want to take this thread off topic so I will make this my last post.

I am always quite willing to make changes for anyone after the 20 minute time span allowed for editing posts has elapsed. Just email or PM me details of the post in question and the changes you would like made.

With regard to smilies, unfortunately all members must be bound by the same rules and I think any more than five smilies in a post is overkill. So use them wisely.

 

[It's Snake Pliskin]

Re: Trading System - Macro Vs Micro

Brett,

Doubtless it has been covered before. And you can build a system which only trades a few stocks (whatever your definition of "few") and is highly optimised for just that basket. But the fewer the securities the less signals that it will generate. This may not be an issue for you depending on your objectives (something you need to ascertion prior to designing a system). You will also need to be right on top of the system to modify it as the securities change character and the system breaks down (the major problem of highly opimised systems). But a high expectancy system with few opportunities will probably not be as profitable as a lower expectancy system with more opportunities.

For example consider a very long-term trend following system which has an expectancy of 2R and generates 10 trades in a year. Your return will be year 20R. But how about a system of 0.4R which generates 120 trades a month? Your expected return could be 48R.

I have two systems, one medium term and one short term and the only price filter I use is > $0.02 so they generate far more trades than I can take which means my opportunity is limited only by capital.

Good post.

It's "fewer".......

 

[Julia]

Re: Trading System - Macro Vs Micro

Good post.

It's "fewer".......

Ah Snake, a pedant after my own heart.
Regards

Julia

 

[brettc4]

Re: Trading System - Macro vs Micro

I realise the system will create fewer opportunities, but as you said MS+Tradesim, you are limited by your capital, so what does it matter it there are 5 opportunities a month or 5,000 if you have this limit.

I dare day, every one has capital constraint so as soon as your system generates a signal to buy you cannot fulfill aren't you effectively not following the system any more? This seems to increase risk as the entries you do now follow are random and does not match what was most likely back tested.

Brett

 

[MS+Tradesim]

Re: Trading System - Macro vs Micro

Julia and Snake...bet you two are just riots to live with :

Brett,

If I cannot take a signal because I have no capital how does that equate with "effectively not following the system any more"? My systems are tested with that factor included.

"This seems to increase risk as the entries you do now follow are random and does not match what was most likely back tested."

Which is why I use Monte Carlo analysis to backtest tens of thousands of possible permutations allowing me to have confidence in the system knowing that irrespective of what signals I take, as long as the live results (expectancy, maxDD, annualised return, R:R, etc) are within the tested bounds then I'm ontrack. I also constantly update and rerun the tests so they include present data concurrent with live trades. I could not trade a system with no back tested/forward tested results as I couldn't trust it. How would I know if the one-off live results were luck or a good system or a bad system?

 

[brettc4]

Re: Trading System - Macro vs Micro

MS+Tradesim, is that level of backtesting available in Metastock only?

I haven't been using it for long, but I haven't seen that sort of sophitication in the Enhanced System Tester. Of course my inexperience in it's use will undoubtably be a factor here.

Brett

 

[MS+Tradesim]

Re: Trading System - Macro vs Micro

Brett,

Metastock's inbuilt tester is a waste of space. I use Metastock to generate trade databases which are then simulated in Tradesim. More info here http://www.compuvision.com.au/

I've never used Amibroker but by posts on here it appears it also has backtesting functions.

 

[brettc4]

Re: Trading System - Macro vs Micro

Thanks MS, Some of their website appears to be down as I cannot see any screen shots etc, and the slideshow doesn't seem to work with Firefox.

Does it use that same language as Metastock or does it use something different?

Brett

 

[MS+Tradesim]

Re: Trading System - Macro vs Micro

Brett,

It uses metastock code and its own additional functions which are easy to learn. All the coding is done in Metastock explorer which uses a dll plugin to create a trade database that Tradesim then uses to run simulations on. All very simple but extremely powerful.

I can view all the website fine in IE. Never used firefox so can't help you there.

 

[nizar]

Re: Trading System - Macro vs Micro

"This seems to increase risk as the entries you do now follow are random and does not match what was most likely back tested."

Which is why I use Monte Carlo analysis to backtest tens of thousands of possible permutations allowing me to have confidence in the system knowing that irrespective of what signals I take, as long as the live results (expectancy, maxDD, annualised return, R:R, etc) are within the tested bounds then I'm ontrack. I also constantly update and rerun the tests so they include present data concurrent with live trades. I could not trade a system with no back tested/forward tested results as I couldn't trust it. How would I know if the one-off live results were luck or a good system or a bad system?

This is textbook stuff!

And Brett, this is the reason why some people run two portfolios of the same system, as returns can be anywhere between the min and max as generated by the monte carlo analysis.

 

[brettc4]

Re: Trading System - Macro vs Micro

To both MS and nizar,

Do you follow a single trading strategy, or do you have different strategies based on price, volume, volatility, sector or anything like that?

Thanks, Brett

 

[MS+Tradesim]

Re: Trading System - Macro vs Micro

Brett,

My medium term (1-3mths) trend following system is based on picking stocks which are outperforming their sector in sectors which are outperforming the all ords. There are a number of conditions involved in the setup and entry but I have a Metastock exploration coded to find them.

I also have a short-term system which will be going live very soon and is designed to take very small bites out of volatility breakouts. Again, Metastock exploration coded to find them. It also uses relative strength comparison as above to pick stocks which offer significant movement probabilities.

And perhaps perversely for a trader, I also take the odd long-term value investment based on fundamentals using very different logic to trading. I enjoy the process of poring over company reports and announcements looking for the undiscovered gems. Gives me something to do with my time as mechanical trading doesn't require much input once all the system development work has been done.

 

[brettc4]

Re: Trading System - Macro vs Micro

Thanks MS,

I think you have answered my original question but in a somewhat different way. Your strategy takes into account the current sector and is therefore narrowing the scope, that bein said, the way the strategy is implemented is consistent across all strategies and therefore macro.

So I guess the answer is, "A little from column A a little from column B".

Your answer has also shown I need alot more practice with Metastock. I wasn't aware you could could preform functions across different stock codes. That being said, I am using free data which does not include information about the sector so did you need ot add that information in, or does your source data contain additional information?

Thanks

 

[MS+Tradesim]

Re: Trading System - Macro vs Micro

Hi Brett,

I use Jose Silva's URSC kit which is a plugin for Metastock and allows me to sort sectors against indices and stocks against sectors. Here for more info-
http://www.metastocktools.com

By the way I get no kickbacks from any of these products. I have just found them extremely valuable and am happy to recommend them.