This way to measure the ordinal association between two measured quantities described by Maurice Kendall (1938, Biometrika, 30 (1–2): 81–89, "A New Measure of Rank Correlation").
In this script I compare Kendall Coefficient and Pearson Coefficient (using built-in "correlation" function).
1) You have two symbols
2) You identified the direction of the first one
3) You decided to buy/sell the first one
4) If you see positive (> 0.7) correlation coefficient then you can apply the same trading action on the second one because correlation coefficients are the measure of the movements relationship of the two symbols.
Both Pearson (built-in script "Correlation Coefficient") and Kendall do the same thing - they measuse linear relationship between two time series.