Chart displays distribution (histogram) of relative deviation of a pair.

Gauss curve

It is possible to compare the relative deviation distribution with Gaussian curve of standard distribution. In general, it can be stated that the closer the relative deviation distribution is to the Gaussian curve of standard distribution, the more stable the pair is. Deviations have a character of a white noise and they will return to normal.

If the distribution is shifted towards one end (ie. it is asymmetric), it means that the pair has a tendency to trend.

Too frequent occurrence of extreme values (above 3.0) means low correlation of pairs components - low robustness.