I have an economic function that is supposed to give a positive result when looking at the market going up, and a negative result when looking at the market going down.
Because this is a general question, and the economic function is very complicated, I will not describe it fully here. However, in brief, the economic function is based on taking various rolling averages of market data over 2500 points, as well as hedge data. E.g. One particular rolling average is calculated by taking 50% of the average of the last 5 numbers and 50% of the average of the last month's numbers. The minimum and maximum of this rolling average are used by the formula.
The function works well in most scenarios. In one particular scenario, the function gives a positive result for both directions of the market.
Now, is there a process whereby I can show that such deviation from expectation is basically a statistical anomaly based on the particular numbers? For instance, if I adjust the rolling average to be 40% and 60% (instead of 50% and 50% as above) then the numbers are indeed positive and negative as expected.
UPDATE: As it turns out, there was a problem with the market data being taken from the wrong place. However I am still interested in the premise behind the question...