January 31, 2012
Recently, in a LinkedIn forum, I presented my latest research note: Optimal Portfolio V. The object was to show that designing holding functions that can increase exponentially in time had a secondary effect on increasing portfolio profitability at an exponential rate as well. Using a stock accumulative process, to which was added a trading component, could produce exponential alpha that went way beyond the Buy & Hold strategy.
I thought that the only way to show my point of view was to apply my trading methods on a stock which resulted in the following IBM chart:
(click to enlarge)
The above chart shows a trading strategy that is pumping cash into the system at the most significant price swings. It demonstrates that by trading over an accumulative process, one can increase performance. Not only by accumulating shares over the long haul but also by accumulating cash as the process grows. Controlling the number of profitable trades could be another way of enhancing performance. It is like showing Optimal Portfolio I, II, III, IV & V in action. The chart shows the last 220 trading days of a 1,500-day test, and it needs the prior 1,280 days to slowly build its inventory and cash reserves.
Long-Term Market View
Having a long-term horizon does not mean that you can not benefit short term (see the IBM chart as an example). The trading procedures partly adapted somehow to the path of the price movement and found ways to trade the short-, mid-, and long-term price swings. All the while accumulating long-term shares. The holding function is exposed to the market 100% of the time, which can be quite different from being 100% exposed.
Any individual or organization seeking long-term investment strategies would benefit from using these methods: retirement and pension funds, mutual funds, hedge funds, and heritage funds. It is not a method that breaks down over time, and neither is it negatively affected by having more and more people using it. On the contrary, the more people use it, the more they all would benefit by indirectly helping each other. The more they all accumulate, the more prices rise on a rising inventory.
To expand on the concept that long-term trading horizons do not mean that there are no short-term benefits, I ran a few stocks using the same script as the one used for the IBM chart with the following results:
( click to enlarge))
All the above charts were produced about two minutes apart; the time needed to run the program and record the output. Each chart shows remarkable trade positioning, and it is not the optimum but a reasonable compromise in time and space for a strategy that does not know what future prices will bring. Some trades are even sprinkled at random with the surprising effect of having relatively good entries. It is not a predictive ability of the program but simply the byproduct of the scaling in function; you will hit some good entry prices right, just by coincidence.
Time & Volume Trade Slicing
Each of the charts shows that risk is being distributed over time by accumulating shares when a long trading window is open (based purely on technicals, not the usual kind, but technicals nonetheless). Some entries are forced (like whatever the technicals on this day, get in), while other entries are more finicky and require to be randomly selected (like if it's your turn today, get in). You even get some entries based on: you have some extra cash on hand so why not buy some more if the other trading procedures do not complain and the buy-window is still open. Technically, this makes the program a sum of fuzzy procedures or, at minimum, an unorthodox set of trading rules.
Whatever the purchases, you won't know what the future will bring, but you do have a stop loss in place to protect you; sometimes soft and other times not so lenient. It will depend on what the procedures were for that particular trade and how prices will behave. But overall, the trading philosophy having been changed to one of accumulating shares, you do know that over time you should do more than well, you should achieve exponential alpha as a byproduct of the trading strategy. One could easily extrapolate from the above charts what will happen next.
Notice also how each stock, without knowing what is coming, tries to scale in and out of positions by partially adapting to the price movement. Each stock has its own signature, and yet the trading procedures are able, on average, to distribute the risk in step with the price swings.
I don't think that any of the above charts are possible using trading alone or a pure Buy & Hold strategy. What was required, however, was for the holding function of the payoff matrix to have a long-term view of the market and that it was ready to accumulate, trade, and hold more...
In my first Alpha Power paper in 2007: I explained that I was using sub-martingales regulated by subordinators as in Lévy processes that could transform an expected zero alpha into an exponential one. It is the same kind of process that is being used in my latest scripts. I am not controlling prices. I am controlling my trading behavior over the entire time spectrum.
And yes, I do over-diversify and use long-term horizons to minimize the impact of stocks that might impair my portfolio. And this way, the losers that I might select will have minimal impact on the overall outcome. I aim for every stock in the portfolio, by trading over the share accumulation process, to produce more profits than would have been possible otherwise. The stocks failing my expectations are simply considered as the cost of doing business, just like commissions, fees, and taxes.
Should Increasing Performance Be Left To Probability?
Wanting to show someone that increasing the propensity to trade would simply increase generated profits and would not be dependent on probabilities required a new test with a higher degree of trade aggressiveness (this required changing only one of several parameters, which individually would also have pushed performance higher). Increasing this one parameter would have for expected outcome to generate higher profits.
( click to enlarge))
The above set of charts trade more on average, 29% more, and generate about 27% more profits. It is not just one chart here and there, it is the whole set that has higher trades numbers. To accomplish the task, a little more of the growing reserves were used to implement all the added trading. But still, the overall cash reserves grew by some 13%, showing that the degree of trade aggressiveness could be increased even higher. Here is the summary of both tests:
(click to enlarge)
Just by increasing the ability to trade more, the program extracted more profits across the board over the entire time spectrum, as if all that was required was turning the trade propensity knob a notch higher.
Published ... January 31, 2012, © Guy R. Fleury. All rights reserved.