From Brett: The following excellent post is from Brandon Wilhite, a full-time trader in the currency markets. While his post is directly relevant to system development and trading, it applies to any form of trading where it's necessary to allocate enough capital to trades to make an acceptable return, but not so much that you run the risk of ruin. He offers his bio below; my comments follow his post.
From Brandon: I have been actively trading various markets for the past two years. I have tried many different markets, types of instruments and trading styles, and have found my niche in retail forex trading. More specifically I am a systematic, but not automated, trader that tends towards longer timeframes and the more out of the way currency crosses. I currently trade my own account for a living. I live in northern Indiana; and my wife is an expert spouse!
Note: The following is easily applicable to discretionary trading with just a little thought.
The way most traders develop their trading systems implies an assumption that once a viable system is found that it will remain viable in the future. In other words, they want to test in such a way to find a system that has traded decently across various instruments and in various types of markets over a long period of time. The end result of this, they hope, is a system that will continue to trade well in the future. While I do not argue with the usefulness of these methods, I do take exception to the final conclusion. Instead, I suggest that the trader does all of this type of testing, but also consciously hold this assumption: “All systems will break at some point in the near or far future.” By “break” I mean that some force(s) external to the system will invalidate it, turning it into a loser. I want to point out that this assumption may or may not be true for a given system, but it is useful to hold regardless of its veracity.
If the trader holds this assumption, then the primary resulting questions would be: “How can I avoid incurring massive losses when this happens, especially if it happens immediately?” and “How can I retain at least some of the profits I’ve accrued before the system starts to break?”
Mainly, in order to assure myself of not being wiped out by an immediate drawdown I like to look at the historical maximum drawdown. I will then allocate an amount of my portfolio to the system that is usually equal to some multiple of the maximum historical drawdown. I like to give ideas time to prove themselves invalid, so I usually use 2 times the maximum. However, I also set up what I call “breakpoints” in my system where I may scale back my size or even put the system on hold. I clearly define all of these breakpoints before beginning to trade the system, and they are usually defined in terms of the maximum historical drawdown, plus logical conditions based upon the market structure that would invalidate the system (for an extreme example, what if the instrument is de-listed?).
I would just like to point out, that due to the relatively large amount of capital I allocate and due to the gearing-down that occurs at my breakpoints, it is a rare thing that I actually end up losing the whole amount that I allocate. There is another technique that can be used to cushion an immediate drawdown, but I would like to move on to the more interesting second question.
Now for the more interesting second question… What I like to do is to determine how much I am allocating per unit that I am trading and work off of this figure. So if the maximum historical drawdown is $1,000/contract, then I am allocating $2,000/contract. I just call this value ‘N.’ To avoid giving back a huge portion of my profits when the system breaks, I will usually multiply this value of ‘N’ by a multiplier of 2 (although a different multiplier can work, depending on how aggressive the trader wants to be, as long as it is greater than 1). When I’ve accrued an amount of profits equal to N, then I will take ½ of the profits and remove them from the system, and increase my trading size by 1 unit. So, in the example above the new ‘N’ would be $4,000. Once the system accrued profits of $4,000 then I would take $2,000 away from the system (i.e. allocate it elsewhere) and increase my trading size by 1 unit. This may seem like overkill on how much to allocate, but if the trader can choose a market that has a very small contract size and then start trading the system with multiple contracts, then the magic of compounding can quickly take hold.
So there you have it, part of a workable solution to managing account growth. This is only a partial solution, because a full solution has to account for portfolio allocation as a whole.
From Brett: One important facet of Brandon's post is the need to distinguish normal drawdowns from abnormal ones. At some point in time, markets change and the methods we've been using won't work as well. That happens to systems traders, and it happens to discretionary ones. By knowing what your normal drawdowns look like--either by investigating your historical performance or that of your system--you have a way of determining if losses are beyond those expectable and normal. That enables you to retain your capital and change what you're doing. By allocating capital to trades that is proportional to one's risk and by growing your size as you achieve success, you are able to accomplish risk control *and* a compounding of performance. One implication of the view that all methods eventually break is that--when you have a superior method--you need to take maximum advantage of it. Brandon has a best practice that accomplishes that purpose, prudently.