Talk:Signal-to-noise statistic

Single source?
The particular formula here seems unusual and the one souce quoted uses the median instead of the mean as stated here. Are there other sources for a terminology defined in this way? Melcombe (talk) 16:24, 23 August 2010 (UTC)

Sum of the deviations or sum of the variances?
It seems strange to add standard deviations in the numerator. The difference of values is a random variable whose variance is the sum of the two variances. Therfore one would expect the denominator to be $$\sqrt{\sigma_1^2 + \sigma_2^2}$$ rather than $$\sigma_1 + \sigma_2$$. --Jorge Stolfi (talk) 23:21, 10 February 2013 (UTC)