Citation

The current study examined the optimal period length using wavelet analysis assuming that TV audience ratings are indicators of field intensity that affects field random number generator (RNG) outputs. As TV programs around the New Year often have high ratings, we focused on these programs in 2012 in Japan. Using Psyleron, Rpgl02, and Orion as physical RNG devices, the sum of squares of RNG outputs during 288 selected programs were decomposed into multiple levels from periods of 250 ms to 256 s through a Haar wavelet filter. Unexpectedly, the wavelet filter could not find sensitive periods, whereas an ANCOVA suggested that Rpgl02 might detect audience rating effects over almost the whole range of wavelengths. Psyleron and Orion devices showed null results. These results suggest that RNG behavior cannot be described by physical signal modeling. There is a possibility of canceling effects in RNG outputs, and this may be a subject for a future study.