Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Random... isn't?

This site may earn commission on affiliate links.
Of the music on my USB stick many songs have never been played, and some songs play quite frequently. Anybody else observing that?

So we started playing songs in alphabetical order, unfortunately the system doesn't remember per-driver position or source (we got his' and her's USB sticks). :-(
 
That's "true random". The system doesn't track what songs have or haven't played -- it simply picks a song at random from your entire library after the current song plays. What it sounds like you're looking for is a less random system -- something that takes your library, randomizes it into a playlist in which each song only plays once, then plays that playlist.
 
It's not true random when songs a, b and c have all played 5 or more times and songs x, y and z have played never. That is called sloppy programming instead.

I know what true random is and that ideally it would track what has played. What I submit though is that it does neither.
 
What each of you want can be found on RANDOM.ORG - True Random Number Service

..and all the theory too.

Suppose you have 100 songs on your memory stick, play them in this order if you like true random:
RANDOM.ORG - Integer Generator
But be warned! You may hear the same song more than once, and others not at all in 100 plays.


Play them in this order if you want to hear all and don't like repeats:
RANDOM.ORG - Sequence Generator
Be satisfied you're getting your money's worth out of all the songs you illegally downloaded.
 
Last edited:
OP, my perspective is similar to yours. My MS "Random" is far less so than what I have in iTunes on my iMac and iPhone. Even with USB Phantom Playing, I tend to hear more of the same songs out of the 6K on my USB stick in my MS, than I do using iTunes with the same wide selection of tracks.

While I could no longer get into details, and I don't want to debate the point, having been a programmer back in the dark ages, I had to write my own random number generator a couple times for applications I was involved with. IIRC, I found various seeds would make the resulting "randomness" more or less so. I simply suspect Tesla is doing something more primative for some reason that ends up selecting particular points in the same sequence more frequently than other algorithms may.

What I do on my MS to sort of get around it, is I sometimes randomly play all the songs on my USB stick, and at others just one Genre or another. My adding that little bit of selection on my own, combined with what Media Player Random does, seems to then let me hear a wider selection.
 
I recall from those programming days too, is that a sample taken from the environment makes for a good random seed.

There are so many sensors on a model S it should be easy to grab a value from one that changes a lot and and run with it.
 
I recall from those programming days too, is that a sample taken from the environment makes for a good random seed.

There are so many sensors on a model S it should be easy to grab a value from one that changes a lot and and run with it.
Very much agree. Things like DOW or HH:MM were never sufficient... Had to get down into dates+time to the milliseconds level to have a real chance of variability over time, or as you suggest, steal some other more precise variable and combine it with others to come up with a seed that is essentially unique. "Random" is just another example IMHO where Media Player works in basic use and had sufficient capability to get out the door, but isn't holding up as well in practical use several years later by a greater number of even more discerning owners. I'm anxious and hopeful to see if Tesla will use 8.0 to address many of these inadequacies (in addition to UI changes and AP improvements), but I'm not holding my breath either. ;)
 
gerti:
When I play my USB in the S I go to browse, songs and randomly choose an alphabet letter and then choose a song.
I am on shuffle so if I choose a song I have not recently heard, it appears that for the following session of songs, that random chooses, well they are those that were not recently played.
 
So, if in 100 rolls your dice never shows you the 5, is that true random or a poorly fabricated dice? As said, I know very well what random playlist generation is and I also know what true random is.
:) I'm sure you do.
This means your dice is more likely to be defected than not, but it is entirely possible that you are just lucky.
Hey someone won the Powerball! That is even less likely than rolling 100 dice without a 5 (about 1 in 82,817,974 vs 1 in 292,201,338). But of course you know this. :) I just put this here in case it is helpful to someone else here!

Although, to be fair, I'd be surprised if they used true RNG in music shuffling logic.
 
  • Like
Reactions: dehydratedH2O
Apple had the "random" problem. People complained and complained that they weren't getting random choices from the iPod, became a Steve jobs level issue. Somewhere back in the GEN 2 days I think.


They hired scientists to look at the randomness generators. It was worth it for other reasons, but in the end everything was definitely decided to be random. People just have a cognitive sense of repetition, pattern matching if you like, even where there is none. So they added a weighted system which lowers the chances a song may be replayed and ups random chance for unplayed songs to hit. Then people thought they had random songs...
 
  • Like
Reactions: dehydratedH2O