Profitable ETF Trading Strategies: applying the laws of large numbers
When you have a statistical edge in the market, the first thing you want to know is how reliable it is. If it turns out that your edge is robust, that means it can be relied upon to work for you in most if not all types of markets. When you have such an edge, the best strategy to adopt is that of a Las Vegas casino. You want to be the house.
Being the house means that you want to play a positive expectancy game with as many iterations as possible, in order to achieve the expected average return of your system.
The statistical edge in the games of chance played in Las Vegas casinos are very small. These small edges though are mathematically certain because of a tightly controlled conditions and the environmental attractions the casinos offer.
Because the conditions are so carefully controlled, the house can’expect for its mathematical edge to work in its favor precisely because of the large number of iterations. If you are the house, you want 1 million people playing blackjack for one dollar hand, rather than one person playing a single hand of blackjack for $1 million.
The results of an individual iteration of a game of chance is not knowable before hand. However, the laws of large numbers work to your advantage when you have the edge.
In the same way, if people in your neighborhood like to gamble, then the best strategy for the whole neighborhood is to pool their money into a single pot and send that money to Las Vegas to play a single hand of blackjack, winner take all. A single iteration at blackjack is the best possible return for your money in Vegas, although it is a slightly negative expectancy gain under most circumstances.
A short-term trader who has plenty of opportunities is better off taking five positions at 1% risk per position than a single position at 5% risk no matter how he decides to rank order the signals by quality. This assumes of course that the signals that pass the screening criteria are equally reliable in the long run.
The greater the number of iterations in the sample size, then the greater your chance of achieving the average expected return a positive expectancy system. There is a natural tendency among traders to try to concentrate their capital on what they consider to be the best trade available. If your system is generating multiple signals though then you are better off taking all of the signals at reduced risk, provided that you have done your back testing work and admin costs of trading as low.