purchased trading systems dont work

seems to be the general gyst of what people on this sectionof the forum are saying....in the med/long term at least.

but i'm guessing that some people here do have succesful mechanical trading systems, and i'm trying to build a fairly simple one myself. the question is to those of you who have developed a succesful system - what basics are your systems based on? did they need a great deal of programming, or nothing that couldnt be done in excel? do you have a level of drawdown before you decide the system is not working? any other info would be useful

thanks in advace

Bowball said:

seems to be the general gyst of what people on this sectionof the forum are saying....in the med/long term at least.

but i'm guessing that some people here do have succesful mechanical trading systems, and i'm trying to build a fairly simple one myself. the question is to those of you who have developed a succesful system - what basics are your systems based on? did they need a great deal of programming, or nothing that couldnt be done in excel? do you have a level of drawdown before you decide the system is not working? any other info would be useful

thanks in advace

Click to expand...

1) Decide what you want from the system BEFORE you begin to code it. You need to look at EVERY ASPECT of it. Your requirements need to match your own personality and your adversion to risk. For example, if you hate drawdowns, then having a system which is potentially profitable but experiences drawdowns around 30-40% is not going to be any good for you. If you need the pychological feeling of being "right" in a trade then developing a system with a low win rate is not going to suit either. The list goes on but you get the gist.
2) Once you have your list of requirements, you can then develop ideas which potentially offer what you are looking for. I would suggest buying suitable software to backtest your ideas. If you intend to trade Equities, then make sure the software has "portfolio" / Monte Carlo backtesting facilities.
3) Make sure you have CLEAN and RELIABLE data on which to develop the system. Rely on free data at your peril !!!!
4) Understand BEFORE you begin this project (i) about the pro's & con's of mechanical systems in general , (ii) limitations of backtesting and maybe (iii) an basic appreciation for Stats. There are a number of very good books on the subject. Not understanding these areas will cause you problems downstream and may make the backtesting stage useless.
5) Don't worry about when to switch the system on & off at this stage. Focus instead on producing a system which (i) matches (or is very close to) your list of requirements and (ii) produces a positive expectancy.
6) Enjoy the Journey.......

I think it's also important to know what type of market conditions your systems was intended to trade in, and when the markets are not acting normally, you turn the system off until the markets return to conditions that the system was intended to work in. Probably the most important lesson I learned...http://www.trade2win.com/boards/images/smilies/icon_cheesygrin.gif

Chorlton has offered very good advice.

As he implied, it is important to understand the trading system development process. This apparent step backward on your learning curve might seem counterproductive because resources will be diverted from researching trading strategies and to understanding a process; nonetheless, such diversion is necessary to properly build and trade automated systems.

I've come to the conclusion that purchased systems don't work because there aren't really any on the market. Ignore auto pilot etc, obviously all cons. I once thought about selling some automated strategies, with full disclosure as to what they do etc, but it's not worth it. There aren't really many people who want a bot which makes 10% a month or whatever, willing to pay a fair price, only people chasing dreams they know aren't really going to come true.

if you has a formula to make money would you sell it?

and you also have to consider market efficency.
every 1 uses a formula that they have purchased, they will not have any advantage.

singh2jatt said:

if you has a formula to make money would you sell it?

Click to expand...

Depends on whether you would trade it yourself... If I stumbled across a system during my own development which had a positive expectancy, etc, etc BUT it did not match my OWN list of requirements (inc my personalility) I may consider selling it.

and you also have to consider market efficency.
every 1 uses a formula that they have purchased, they will not have any advantage

Click to expand...

To some extent this is true, as over time the system would lose its edge due to increased slippage, etc. However, if the details of the system were made available (unlike a blackbox system) then it could be "tweaked" to make it more unique.