Period - This is the number of years of historical data that will be used for calculating correlation coefficients. Correlations can be calculated over various time periods ranging from two years to 40 years. The longer the time period selected, the fewer results will be available because many commodities and stocks simply do not have long histories. However, longer time periods may yield more meaningful results. Your results may show a shorter time period than requested, if the desired period does not exist for the selected market or markets. The period used will be disclosed on your results.
We strongly recommend against using short time periods. Statistical significance reflects, unfortunately, only the likelihood of a false positive. This analysis sorts through millions and millions of pairs, and reports the more promising pairs. Using more history gives us better confidence in the correlation coefficients, which reduces the number of false positives. Reducing the meaningless, coincident relationships reduces the effort needed to avoid mere coincidence. No amount of data will provide certainty, but to start with the odds in your favor, use at least 15 years of history. If you wish to trade a market that has existed less than 15 years or an instrument that has changed its character, you may be forced to use a shorter time period. Be advised that false positives are more frequent with such a data set.
Close Window [X]