Forum menu
Probability
 

[Closed] Probability

Posts: 5169
Free Member
Topic starter
 

I was always useless at maths.
I have 1000 songs on my playlist. I randomly play 100 songs a month (this can include the song being played more than once).
What is probability that I will not play one particular song at all during the year?


 
Posted : 28/01/2020 10:32 pm
Posts: 17313
Free Member
 

Nil, because the shuffling algorithm isn’t really random


 
Posted : 28/01/2020 10:35 pm
Posts: 8527
Free Member
 

Just 3 weeks ago, when sitting maths exam at uni, I could have gave you an answer.

It's now gone.

'S how I roll.


 
Posted : 28/01/2020 10:38 pm
Posts: 1320
Full Member
 

If the algorithm is truly random, then you have a 90% chance of not playing it each month.

0.9 x 0.9 x 0.9 twelve times.= 0.3 ish so 30% chance of not playing it in a year.


 
Posted : 28/01/2020 10:39 pm
Posts: 1340
Free Member
 

Also nil if the particular track is something your eight year old daughter put on there.

Matt


 
Posted : 28/01/2020 10:40 pm
Posts: 1340
Free Member
 

It's higher than that smudger as you could play the same track multiple times each month. My guess is .999 times .999, 100 times, then multiply by twelve for each month.

Matt


 
Posted : 28/01/2020 10:41 pm
Posts: 6089
Full Member
 

In theory, the chances of a track not being played in a month is 9/10.
The chances of a track not being played in 12 months is 9/10^ 12, or about 28%, apparently.

Assuming the shuffle is entirely random, of course.


 
Posted : 28/01/2020 10:42 pm
Posts: 207
Full Member
 

Smudger, You're assuming it plays 100 different songs per month. Surely it's 0.999 per play rather than 0.9 per month...

In fact, I think the months are irrelevant. It's 1200 random songs per year so 0.999^1200 = 30.1%

Might be wrong though....


 
Posted : 28/01/2020 10:43 pm
Posts: 1320
Full Member
 

Yup - missed the bit that they can be repeated in a month.


 
Posted : 28/01/2020 10:45 pm
Posts: 300
Free Member
 

If I remember correctly:
When sample n times from the set {1,…,x}, then the expected number of unique values is x[1−(1−1/x)^n]

In one month the number of unique songs should be 95.207852886. So 4.8ish repeats.

In 12 months I think: 698.986570907 unique songs - 302 ish repeats.

I'm too old for this stuff!

#Edit - the 'power of n' in the formula got lost


 
Posted : 28/01/2020 11:01 pm
Posts: 8836
Full Member
 

In fact, I think the months are irrelevant. It’s 1200 random songs per year so 0.999^1200 = 30.1%

Gets my vote


 
Posted : 28/01/2020 11:22 pm
Posts: 3273
Free Member
 

What is probability that I will not play one particular song at all during the year?

Answer = unlikely.

But it's also possible the same song could repeatedly play non-stop for the next year.

Just don't walk down Easy Street.


 
Posted : 28/01/2020 11:31 pm
Posts: 9193
Full Member
 

If this is a Spotify playlist, my experience suggests that it is considerably more likely to play songs at the end of your playlist than at the beginning. I haven't heard some of the ones near the start for ages.


 
Posted : 29/01/2020 12:55 am
Posts: 28593
Free Member
 

Nil, because the shuffling algorithm isn’t really random

None, because Spotify's shuffling algorithm doesn't work properly.

Just don’t walk down Easy Street.

*Shoots Glenn with crossbow*


 
Posted : 29/01/2020 8:53 am