strategies, indicators and so forth

[JayRemy]

Hi Guys,

I've been messing around with the tradestation easy language stuff these last few days, mostly strategies seeing what happens with different conditions and how profitiable certain systems are - it seems to be the quickest way to back test something!

If anybody wanted to suggest some system, theory of a new indicator i'd be happy to code it up and produce some adhoc stuff to see what we come up with.

Its really caught my interest at the moment and it would be great fun to produce something cool.

Cheers guys, Jay.

[torero]

Welcome to system trading world, jay! Enjoy your ride! It's a new interesting, fascinating world... but it's also lots of headaches and failures as well if you are not too careful with it.

I do many analyses of indicators and systems and backtested them in many timeframes and symbols. Most fail... except for a few. But those are keepers. This is also the reason I came to a conclusion that most indicators if not all are not useful and unnecessary in trading except price. But don't get me wrong. There are those who them use in their trading with good results. I myself just haven't found it.

[JayRemy]

hey torero, thanks man :p

how do you know if youve got a good system?

silly question maybe! 'cos a good system makes money, right?

I've done this little strategy on a 30 minute EURUSD, which over the past 2 years has a profit factor of 1.76, and so seems like it could make a stack of money. It looks pretty good to me, but i dont want to get too excited too soon, incase what i've done is just trash and wouldnt work in the real world.

Theres some stuff which I'm unsure of, is there any kind of general rule for slippage?

Infact, going through this process of trying to create a strategy and backtesting it has taught me alot about trading in general. I can clearly see money management i.e. stop losses, scaling out profits etc are so important it can turn a non-profitable trader to be profitable.

[torero]

Once you're done you need to walk-forward it (watching it in real time) for a few months and compare to see if the results is similar to what it's been tested to historical data. Most of the time, strats fail in the real time due to not considering adding other factors in such as slippage and commissions, etc. If it does well in real time after a few month, let it go with real money with small lots and see how it goes. Need to keep an eye on the Max drawdown. This is the biggest problem for many traders. The guide is to have max DD x 2 the money to cover these dips. You also have to decide where the cutoff point is (pulling the plug when it reaches a loss). Many use the trendline of the equity curve or 2x max DD for example.

Adding slippage and commission into the strategy testing makes the difference between heavy losses and extreme gains.

Without commissions and slippages



With commissions and slippages

[JayRemy]

thanks for that torero. i'm going to walk with it for a while, see how it performs. although, i've got a scary max draw down figure. it's coming from a period of time right in the middle of chart and doesnt happen again but i'll sort out that problem before market open hopefully

thanks for the advise, cheers Jay.