Q: I feel that I have a good handle on the concepts of limiting risk and position sizing as they apply to trend following strategies. However, I have a bit of a conflict when it comes to mean reversion trading strategies.
My objective for the mean reversion strategy that I trade is to look for ETFs that are in a long term trend and have deviated from that trend on a short term basis. For example, a particular ETF may clearly be trending upwards on a long term basis but is experiencing oversold conditions on a short term basis. The short term pullbacks have proven to be a good entry point for short term trading. If the position moves against me, that means the ETF would have become even more oversold. As long as the initial conditions for an upwards trend are in place, this would be an even more advantageous point for me to enter the trade, so I would typically add to my position. This seems like a paradox, but it has proven successful.
I limit myself to no more than 3 or 4 entries due to increasingly oversold conditions, and I typically pyramid the size of my position as the ETF becomes more oversold (for example, I may buy 100 shares on my first entry, then add 200 shares as it becomes more oversold, and finally add an additional 400 shares for my third and final entry.)
This method has given me both a positive expectancy and a high probability that the trade will be successful (often in excess of 80% correct.) I exit the trade either when it is no longer oversold on a short term basis (usually locking in a profit) or when the ETF is no longer giving me the signal of a long term upward trend. This serves effectively as my stop loss. It has happened, but it is rare.
Unfortunately, since I look for strongly trending ETFs, they need to drop a lot before I no longer get a signal that they are trending. If I use this basis as my maximum R and limit my maximum risk to no more than 1% or 2% of my trading account, I end up with exceptionally small position sizes which are not worth trading for the small profits that they typically yield. Using smaller stop losses at arbitrary points does not improve my expectancy, and it is in conflict with the philosophy that as long as the basic long term trend is in place I can typically expect it to revert after a pullback.
What is your advice for position sizing in a case like this when using a mean reversion strategy? Thanks.
A: You didn’t give me enough information to provide you with valuable position sizing advice, so what if I give you some things to think about?
As I say frequently in the Definitive Guide to Position Sizing, your objectives determine your position sizing strategy. What are your objectives? You did not say. Are they clear and fully defined? Have you put a lot of effort into thinking about them? Crafting a position sizing strategy is much easier after you have done enough work on your objectives.
As for a scaling-in strategy, first determine the maximum risk per position you are willing to accept. Then divide that into the maximum number of times you want to scale-in. Something else to consider: what happens if you have max positions on and the market type changes from bull to bear and you start losing 50%+?—Van
