The n-day SNR for a given market is calculated by taking the absolute price change over an n-day period and dividing it by the average n-day volatility.
SNR(HLC, n, ...)Object that is coercible to xts or matrix and contains High-Low-Close prices.
Number of periods for moving average.
Other arguments to be passed to ATR.
A object of the same class as HLC or a matrix (if try.xts fails) containing the signal to noise ratio.
$$SNR_n = \frac{|C_t - C_{t-n}|}{ATR_n} $$
Using average true range as the volatility measure captures more of the intraday and overnight volatility in a way that a measurement of Close-to-Close price change does not.
The interpretation is then relatively intuitive: an SNR value of five indicates that the market has moved five times the volatility (average true range) over the given look-back period.
Skeggs, James and Hill, Alex (2015). Back in Black Part 2: The Opportunity Set for Trend Following.