When you start running Monte Carlo simulations of trading strategies it soon becomes evident that every unprofitable strategy can eventually yield some sort of shortterm profitable result. This is the inevitable consequence of the probability of getting a profit always being higher than zero and it manifests itself as the appearance of “profitable batches” along any system’s equity curve. If every unprofitable system (with no edge or a negative edge) will eventually get some profit then it becomes a very relevant question to ask if the results of our profitable systems are indeed due to a true edge or merely the result of a temporal profitable outcome of an overall losing strategy. Through the following paragraphs I will discuss this problem and show you how Monte Carlo simulations offer us a very powerful tool to evaluate the like hood of our strategy’s results being just some “randomly profitable outcome” of a losing system.
Whenever you flip a coin you can either get heads or tails, everybody knows that if you started to bet with someone each time you flipped the coin you would eventually just break even (as the number of bets tends to a very large number) but along a short number of cases you can either win or lose a lot as “streaks” of profitability or losses tend to happen asymmetrically from time to time. If your friend was also charging you for each time you would want to bet the outcome would even become worse (inevitable net losses in the long term) but the possibility to see short term profits still remains.
–

This is a very important characteristic of trading any market. Whenever you trade you have a probability to come out as a winner or loser on each trade and even a strategy which makes no sense and will generate a wipeout in the long term will produce some shortterm profitable results. This is – without a doubt – one of the largest problems new traders face because they tend to trade strategies that work for the “short term” without any idea of whether or not they can expect long term outcomes to be successful.
The question – in the end – is to know whether the outcomes of our strategy can be expected to be the result of a net losing noedge strategy (which has a 50% chance of being a winner or a loser but loses in the long term due to trading costs) and with what probability this might be the case. If a strategy – for example – took 500 trades within a 10 year period we could run a 100K iteration Monte Carlo simulation of a netlosing strategy for this number of trades and estimate what the probability would be to achieve a certain net overall profitability.
After running the above Monte Carlo simulation it becomes evident that a strategy with these characteristics (risking 1% per trade) could reach profitable outcomes with a quite high probability. For example, the probability to reach a 20% profit within a 500 trade period for a random netlosing strategy is 57% showing that arriving at such a profit during the test does not guarantee that the strategy has an edge. However if we go up to a 300% profit (three times the initial deposit) we get that the probability to reach this through a random outcome is 0.032%. Showing that very high profit levels can STILL be achieved through a random strategy although with a very small probability.
This analysis shows us that most strategies have a certain probability to just be profitable due to random chance but we could safely say that the probability of this being the case is acceptable whenever it falls below a certain threshold (for example 5 or 1%). However performing this analysis is vital to determine both the statistical significance of the results and whether or not the results we have can be said to be meaningful when compared to “trading randomly”. Many people would just tell you that a “3 month test” is enough to say whether a system “works” but they simply ignore that this testing period is too short and a netlosing strategy can generate very profitable outcomes during this time just merely due to random profits.
This weekend on Asirikuy I will be releasing a video better explaining the “against netlosing case” evaluation, showing how this can be performed using our Asirikuy Monte Carlo simulator and what different aspects beyond the probability to reach a certain profit or loss might be taken into account to judge the probability of a system’s results to be the outcome of a netlosing strategy. However I hope the above article makes it clear that profitable outcomes are NOT merely the result of profitable strategies and profitability can in fact be reached – up to a certain point – through random profits. This is why it makes no sense to just setup an arbitrary criteria for short term system evaluation but evaluation time needs to be aligned with the amount of trades necessary to prove profitability beyond randomness.
If you would like to learn more about Monte Carlo simulations and how you can use them to better understand the long term statistical behavior of a strategy please consider joining Asirikuy.com, a website filled with educational videos, trading systems, development and a sound, honest and transparent approach towards automated trading in general . I hope you enjoyed this article ! :o)