The Probability is 1 in 589

[tomcat]

So I have my winamp playlist with with 590 songs. It is set to shuffle. I see the next song I want to play, but before I have a change to click on it my current song ends and winamp shuffles to a random track way down in the playlist. Too lazy to scroll back up and select my song, I press the next track button. And what do you know. Winamp goes right to the song I wanted. Cooool Way to go nullsoft.

Just wanted to share. The song was DDT - Rodina.mp3, excellent one for those that understand russian.

Ok that was my random post for the day, better go do bio

 

[deftron]

What numbers should I play in tonight's Pick 3 ?

 

[Pretender]

I believe it was mentioned a while back that Winamp's "shuffle" option isn't truely random - it favors the song which you listen to more often more. I've seemed to notice this when the songs I clicked on quite a bit a few days ago came up more often (considering the playlist was 400 songs long, I doubt it was simple coincedence). Think about it.

 

[Wagner]

Actually, the probability could be as high as 1 in 1. The algorithm works as follows: If you have 590 songs queued in your list, Winamp plays an arbitrary song from that list. The next song is chosen from any one of the remaining 589. Once the next song is done, the song after that is chosen from the remaining 588, and so on. The more songs you play, the higher the chances of hearing a song that you have not played yet.

 

[tomcat]

Wagner, that is an interesting point. I was thinking about asking how truly random winamp is, I believe that I played only 2 songs in the playlist before my streak of luck struck. Interesting. While on the subject, can pcs truly do anything random? I think there are random number generator chips on motherboards and calculators, is that correct? Do they also follow some kind of algorithm?

 

[piku]

I read before that its impossible to compute truly random numbers, but they may have fixed that with special chips or something.

 

[Pretender]

It is impossible to generate a truely-random number. The pseudo-random numbers generated by computers are adequately "random" so that we won't be able to easily predict the next number generated.

 

[Wagner]

It is impossible to compute truly random numbers on deterministic automata (namely any binary computer, calculator, electrical circuit, etc). Deterministic automata are by definition deterministic, namely predictable ahead of time. The best we can do is generate pseudorandom numbers, namely seed extremely complicated algorithms with numbers that are extremely difficult to replicate. (Say, the miliseconds component of the current time.) But this is still deterministic, and can be replicated by an adversary with access to the necessary information through legitimate or illegitimate means. Quantum computers, when they do become mainstream, will allow true random number generation.

Until then, you can get your random numbers from hotbits, which generates random bits based on the decay of radioactive matter.

 

[yakko]

The chances are 50/50. It is either yes or no.

 

[tomcat]

interesting, thanks for the reply guys. I remmember seeing a slashdot article a while back where they were converting a video stream into ascii and then using that to generate random output. Seemed interesing, also would be cool to see what the quality of an ascii video feed would be, if the conversion is possible in realtime it should be possible to do some low bandwith video conferencing.

 

[GreenBeret]

yakko: No. That's wrong. If it is 'random' then then chance is not 50/50 *sigh*

 

[yakko]

GreenBeret,

It is 50/50 since there are only two options. Either the song will come up or it won't. Yes or no.

 

[Wagner]

tomcat: A video stream converted to ASCII would also provide appropriate (difficult-to-predict) seed values for a pseudorandom number generation algorithm.

If you're interested in ASCII video, there's a NES emulator that produces ASCII video output. It's called TextNES.

 

[StageLeft]

yakko Thats like saying that if I buy a lottery ticket for $10 million I have a 50/50 because either i'll win or I won't. It totally denies statistics.

 

[Wagner]

Yakko,

Let's use your logic.

Within the next 60 seconds, you will either die of a lightning strike, or you won't. So the chances of you dying of a lightning strike within the next 60 seconds is fifty percent.

Combinatorics is so simple! If I only knew, I wouldn't waste my college career studying it!

 

[StageLeft]

Here's one - in the next hour I'm either going to win $1million or $2 million or neither. Thats a 33/33/33. Not bad

 

[Wagner]

Skoorb,

With odds like that, your expected winnings are precisely one million dollars. Can I be your new best friend?

 

[StageLeft]

If I win there is a 50/50 chance i'll give you all of it

 

[yakko]

I see you guys understand.

 

[Muadib]

Note to self, buy lottery ticket...

 

Mabey you have a connection with yout computer at a higher level, and you willed it to happen.
Probability does not exist

And why do you have a playlist of 589 songs?
Why not just play the ones you want, when you want.

 

[Wagner]

How on earth is your computer supposed to know which ones are your favorite songs? Isn't the thing on random play?

What's happening is that you are tending to remember instances when your favorite songs play. The songs that are not necessarily your favorites are appearing to be background or ambient noise, therefore not making as significant an impression in your mind.

 

[Pretender]

<< Hate to say it, but Pretender is right on this one. >>

I feel I should be insulted, but you did agree with me

Anyway, I have no &quot;proof&quot; behind my reasoning, but I've noticed that somehow, the songs I allow to play longer (since my playlist consists of my entire MP3 directory, the songs I dislike I tend to quickly skip) or click on to play tend to come up more often later on. How it does it, I don't know, maybe it stores the information in a temporary file somewhere or in the registry, but I've never searched to find out.

In a quick google search, the only thing I found was something on Winamp's site, if you look down at Winamp 2.24 updates, you see something briefly mentioned as &quot;Better shuffle logic&quot;. Granted, given the lack of details, this could mean anything, but it could mean that either they improved it to favor songs more preferable to the listener.


I have no proof of anything I'm suspecting, and there's no way to prove it other than by looking in Winamp's source code or running an experiment, and considering as how the first choice would be near impossible, and I'm too lazy to do the second, it'll just be a mystery

 

[tomcat]

<< I'm too lazy to do the second >> LOL@Pretender

 

[LordMaul]

QUICK! Go buy a lottoo ticket!