So I came home and used an album with 21 tracks as my sample. Using Track 10 as the starting track, I moved on to the next track and took note of the frequencies. I planned to take 100 readings but gave up at 50. Assuming that the track selection is perfectly random, the probability of any single track should be 0.05(1/20), which equates to 2.5 times per track. My findings below show otherwise.
This obviously shows that the iPod's algorithm for shuffling songs is not concrete enough. Note that Track 11 was played a grand total of zero times. Now we all know why some songs are not played at all, even in shuffle.
4 comments:
Yo, I spy in the options panel that there might be an option for more randomness vs random within a artiste. This might be the side-effect of the biased algorithm.
I am aware of that. But mine is set to middle indicating "Random"
Yep; but the fact that this slider exists would indicate the possible existence of the problem that you'd highlighted.
Good observation though. =)
Hahas. Thanks. Shows that "random" in iTunes is anything but.
Who are you by the way?
Post a Comment