Step 1
I’ve explained what an anchor is and how I do consider its price trend to have predictive value. Gold is very dependent on currency trends but diverges frequently enough that it’s better to also take its price into consideration.
Step 2
By checking the trend of another associated market I can reduce false signals. I don’t expect the other market to (ever) provide a lead – just a screen to improve my hit ratio. In the case of gold there are a few obvious markets to check: The Euro, The Swiss Franc, The CRB (or GSCI) index, and Oil. It turns out that Silver works best.
Step 3
I have to then choose an indicator to improve my addons so I can trade in and out of Gold during times of trendlessness and increase my size if I get an appropriate pullback. I always use the same oscillator but in the case of an anchor I allow the machine to choose between that indicator as a function of Gold or Silver. It chose Gold.
Structure
Even though the system follows the trend it is set up to expect and trade mean reviersion orders within the trend. We have to survive the chop! Keep in mind, the oscillator I use takes trend into account.
Here’s an image of how the system trades:
The real test to see if my assertions are valid is to look at the theoretical performance of the system set up to use these variables. Walk forwards are aggressively used (no overlapping periods, 2008 is excluded from the sample periods). Here’s the equity chart for the 2003 – current period:
This curve gives me confidence when it comes to deciding if I want to put on a gold position and when it is suitable to put it in a pair trade with a correlated contract like the Euro.
