• MP3 Players
• Stories
• iTunes: Just how random is random?

iTunes: Just how random is random?

By David Braue on 08 March 2007

• 

  Tweak commented on 22/04/2009 19:53 Report abuse

  4068 songs on my pod and shuffle just played 3 Amy Winehouse songs in a row... Shite!

• 

  SVS-NS commented on 18/04/2009 23:43 Report abuse

  Hi,
  
  I have enjoyed this discussion very much.
  
  This has been bugging me since I first started using iTunes. I can't comment on the statistical analysis, but 8 years of continuous use does make one notice certain patterns.
  
  It does indeed seem that it is difficult to override the "most played" feature. The playlist can get stuck in patterns that seem hard to break. This can be frustrating.
  
  If I happened to listen to a song ten times yesterday, and I switch to "shuffle" today, I might definitely NOT want yesterday's favourite to be weighted heavily in the "random" playlist—but there it is, whether I want it or not.
  
  I do realize that in my case, 1000 songs is a small sample, but I must say that there are songs that have NEVER ONCE come up in 8 solid years of shuffle play, and others that come up with far too much frequency . . . which in turn seems to make shuffle pick them again, and so on . . . and this seems to be the source of these questions about "randomness."
  
  Also, there are always a few coincidences that seem far to "cute" and "apple-y"—like any time a Leonard Cohen song shuffles up, many, many times it will be followed by Melys' "When You Put Leonard Cohen On." Kind of funny the first time, less funny the fiftieth time. There are a few other combinations that do have the feeling of being "programmed" rather than coincidence.
  
  Always intersting to read about the mysterious inner workings of these things that have become such a ubiquitous part of our lives.

• 

  Mark commented on 12/04/2009 11:11 Report abuse

  I find it very amusing the volume of comments here who think they understand statistics. I'm a professor of statistics at BYU and most of the comments here are idiotic. @A Fangil - the definition of a random variable is "A FUNCTION with assigns a number to an outcome" - to say that random is by definition without a formula or pattern is ridiculous. All "random" events have a pattern and function that defines them. Take a coin flip. It follows a binomial distribution. Take disease survival, it likely follows some sort of Poisson distribution.
  
  While it is true that this article doesn't use statistics to justify it's findings, what is ridiculous is to assume that the sample size is too small to make conclusions. And to the person who said that statistics would only tell us the probability of the results being due to chance - well, duh! That's what ALL statistics do. That's the point, if our observed outcome only has a 3% chance of being due to random chance assume our null hypothesis, that could be strong evidence that it WASN'T due to chance. Especially if you repeat the study a second time and get similar results.
  
  But still, the sample size is likely fine (did anybody bother to check the expected counts if you're thinking Chi-Square is the way to go?) and NO you don't need to repeat the whole process several times to make your findings legit, you need to make sure you have more than "one subject per treatment" just like they did. There's more than one "top hit" more than one 'randomly' selected artist with 5 songs and there's more than one MP3 song. You don't even need to have each song represented in AAC and MP3 form, ever hear of incomplete block designs for experiments? While the method for selecting artists and songs could have been better, it does seem relatively random.

• 

  Hillio commented on 13/03/2009 13:16 Report abuse

  I think they removed the randomization "slider" from the latest version of iTunes too...at least I can't find it.

• 

  mynameisgabe commented on 29/01/2009 08:20 Report abuse

  I listen to iTunes all day on my computer at work, on the iPod on the drive to and from work and on the appleTV at home and I definitely see the same songs on all three devices played more frequently than others. It's really frustrating.

• 

  madbeppo commented on 17/01/2009 07:11 Report abuse

  I have over 11,000 items on my ipod. By far the greater number are from CDs I have ripped; and by far the greater number are classical. In addition to puzzling near-repeats (same album, same composer) when in shuffle mode, ipod also seems pretty much to ignore albums that were ripped earlier. I could understand if, when shuffling, ipod (or itunes) first made a (random) selection from the whole collection, then shuffled amongst this selection. But why does it seem never to choose albums that were put on earlier? --On my laptop I have an earlier version of my itunes library. Shuffle comes up with a very different selection of music than it does with my complete collection!

• 

  tfp commented on 05/01/2009 03:50 Report abuse

  It may be merely conjecture but it confirms what I, and others, have noticed while listening to hours of "random" shuffling from iTunes. There is something at play here, and its not just the music.

• 

  heyjohnny commented on 03/01/2009 21:08 Report abuse

  There's no way it's merely by chance. I have 3 Motorhead albums on my itunes and when I generate a playlist it's always the same 2 or 3 Motorhead songs that are played. There's definitely something off about itunes shuffle algorithm

• 

  A Fangirl commented on 01/01/2009 18:56 Report abuse

  I agree with everyone who's saying that this article is not proving anything. Random in its literal sense is BY DEFINITION, not having a pattern or formula. While I do agree that it can be annoying hearing the same songs over and over, that's merely by chance. So, all of you people complaining that you want the iPod to look at your playcount for each song, pick new songs, etc is going against the whole point of random.
  
  Doing one trial is not enough to shout CONSPIRACY! As someone else said, if you flip a coin 1000 times and all get heads (which is possible. Just because there's a 50-50 chance doesn't mean that there HAS to be half heads and half tails), it doesn't mean it's not fair.
  
  Plus, I think there is some psychology to this. If you start thinking that iTunes isn't random, you're going to start looking for patterns. (Which is normal. As humans, we look for patterns in randomness.) And if this is the case, you are focusing on the ones you think you're hearing more than others, when you may be actually hearing a bunch of songs that you haven't heard. Considering the author(s) went into this article already feeling there was a conspiracy or at least that iTunes wasn't that random, of course you'd be looking for examples to prove your point

• 

  musicyoda commented on 04/12/2008 06:30 Report abuse

  YouTube Random Music Video gadget I use from google gadget is far more better in radomizing music list.