Equity Curve Trading

[Kiwi 16]

I have always thought that trading ones equity curve was the nastiest form of curve fitting.

 

Until this weekend.

 

 

I have an excellent system. It captures the majority of the trend on trend days. Yep. Almost all of it. And its got a great profit factor in my in sample test. Really simple TA with next to no optimization. 1-2 trades per day per market (avg 1).

 

Wow. Yes. The holy grail.

 

And then I tested it out of sample. Oh ... there's a nasty hole there ... and I can't see what I've done wrong. So I took a swing for the dark side and applied a couple of simple equity curve rules ... and the holes became small dips and the world became warm and fuzzy again. Profit factor doubled for OOS test.

 

Now, I will reveal a little more later but can say now that I can't detect the dip before it happens and the equity curve is the quickest way of spotting it .. and coming out the far end of it.

 

So, wise friends, what do you think about this? Is it the blackest of witchcraft or is it a legitimate technique? Any ideas or pointers on how to make it most efficacious?

[gooni 10]

Henry Carstens from Vertical Solutions will trade something similar, relating to the 'control graphs' of his systems. While this is not equity curve trading, it is turning the system on and off based on hindsight, which is similar enough (but probably more sophisticated). He has pedigree as an automated trader, so at least the concept (if not the method) is sound.

 

Malcolm Morley has spoken about his position sizing in the past. Variable Fraction Sizing is what he recommends, and that naturally scales in during the good periods, and out during the bad.

 

The question that needs to be asked, however, is whether:

 

1. The edge is persistent enough that the 'good' periods outweigh and outlast the 'dips'?

2. Is there a strategy that would benefit from trading during only the 'dips' but not during the good period?

 

To me this would suggest that you've potentially determined a method for systematically deciding on the market condition, and there is no reason why that shouldn't be exploitable in a way.

 

Regarding curve-fitting - well, that depends on the complexity of your strategy :-) If your degrees of freedom are small already, adding another (equity curve moving average or whatever) is not a big deal.

[brownsfan019 17]

Interesting post Kiwi. I would say that you possibly found a way to minimize the drawdowns, which is obviously nice. The concern is whether or not that you found a blip that helped in hindsight or a true piece of the puzzle.

 

From there it's incredibly difficult to know whether or not you have your grail until it's tested live in current conditions. If it performs well on the right-side of the chart for what you consider to be a good testing time, then you might have it.

 

Personally, I've found that pushing my trading over to the consistently profitable side wasn't changing setups or finding more ways to trade different ways; rather, put in risk management rules that shut off the bad days/markets and keep me going in the good ones. IMO that's more powerful than constructing the trading plan. I honestly believe that just about any trading plan out there could make money if a good risk management policy was in place - whether that be take 1 trade and done, trade all day, only trade the AM session, etc. etc.

 

I don't think of this talk as taboo or anything, quite the contrary actually... so much emphasis is placed on trade entries and exits, very little is discussed on ways to manage your risk and drawdowns. So much is placed on finding the perfect entry or exit and very little discussions about entering on a simple pattern and then using risk management techniques to maximize the gains and minimize the losses. In theory if one could find the right combo, you could apply it all over the place in a variety markets cranking out profits. One day that's where I hope to be at least... turn on the ATM and watch it print money.

 

Well done and keep us posted. If you want to talk in further depth and would prefer it off the forum, feel free to PM me. I know there's a fine line between wanting help and offering help on a free forum vs. keeping the money makers to yourself.