Find a strategy. Validate It. Do It. (Part 2)

Last week we looked at the BORD pattern traded on the S&P 500 ETF ($SPY). It offered an elegantly simple setup with reasonable results. As presented, this is known as a ‘single market system’ meaning it was designed and built to trade one market. However, single market systems tend to be based on random noise (found through data mining or optimization) rather than a specific edge. As such they ultimately fail; what was once a profitable pattern deteriorates to the point of becoming unprofitable.

We’ll discuss this in more depth in a later newsletter, but for now we want to put the BORD pattern through a simple exercise to see if it does actually have any merit as a robust setup. If it does then it should be profitable across a larger universe of stocks and not just on the single ETF we’ve presented. To do so we’ll test it across the entire S&P 500 universe including all delisted stocks to remove survivorship bias, using a maximum of 20 positions with 5% of capital allocated to each. The test also includes commissions.

Source: Premium Data
Whilst equity does increase, growth is a little lumpy with a few stiff drawdowns. It certainly appears better than a random pattern, but it’s not overly appealing from a trading standpoint. Next we’ll add an index filter ensuring that we’re always trading in the direction of the prevailing broader market trend, in this case we’ll use a 100-day simple moving average of the S&P 500.

Source: Premium Data
The equity growth here is certainly smoother but there is a definitive flat period between 1999 and 2003 with no growth at all. This was the major bear market after the tech crash so the index filter has kept the strategy sidelined for long stretches. Not many traders will withstand an extended period of no profits.

Lastly we’ll also apply a 100-day simple moving average filter to the individual stock itself. So, we want the BORD pattern to be in a bullish broader market environment (S&P 500 trend up) and we want the individual stock to also be in an uptrend.

Source: Premium Data
These last results offer the smoothest equity growth, but profits have diluted by 50%. This is a common trait of filtering – it might make the trading process more comfortable but comes at the expense of profitability.

In summary, the BORD pattern appears quite robust but needs to be put through more rigorous testing and simulations. Though this setup is probably not overly attractive on its own, it could be one setup within a toolbox of various robust setups which, when combined, offer a significant trading system.