Reminds me of the old iTunes shuffle thing. When it was first introduced it was actually random but too many people complained it was broken when they heard the same artist multiple times in a row so they rewrote it as a shuffle algorithm that would feel more random than actual random.
Just goes to show, we don’t actually want random, we want variety.
There are two basic ways of doing “random” songs:
- Pick a new song randomly each time a song ends. This is the naive way to do it and can result in playing the same song twice.
- Randomly shuffle the list of songs once and then go through the shuffled list in order, guaranteeing that no single song gets played a second time before all songs have been played.
The strategies are different, but I’d argue that they’re equally “random.”
I’ve got a cheap Chinese aftermarket head unit in my car that uses strategy #1, and it’s mildly infuriating.
Yeah, but all modern music platforms use a more advanced random, where it will avoid putting two different songs by the same artist in a row for example. But it’s still based on the second strategy you wrote.
This seems somewhat flawed. Lets say you have 90 songs by Vengaboys, and 10 songs by Slayer in your playlist. In order to play every song without playing Vengaboys back to back, you’d need to play Slayer 4x more often than you play Vengaboys.
deleted by creator
Reminds me of an article I read long time ago of the need in computer games to tweak percentage chance of success and failure, because if it is true as presented 80% success rate players think it should be “almost always” and complain when one fifth of attempts fail.
Me when the weather app says 80% chance of rain, so I go everywhere with an umbrella but it’s overcast all day long. Then it says 15% and I get rained on while walking to the store.
I’m never going to learn.
I found out recently that those percentages actually mean 80% of the local AREA would have rain and 20% would not. Meaning if there is a chance of rain in your town at all it’s likely raining somewhere even if it’s just a tiny drop or two.
So if you don’t want to get wet at all bring an umbrella if the chance is over 0%.
Source: was talking to a meteorologist about this exact thing.
Well rain chance is a compound probability it’s the probability that it will or will not rain multiplied by the percent of land hit with rain. Like if 50% of an area will be hit and there’s an 80% chance it will rain the number the weather Channel will give you is 40%
Chance of rain is calculated based on two things: Meteorologists’ educated guess on the chance of rain, multiplied by the area that will receive rain, in the event it does rain.
Pokemon players are very versed in this. This is why I only use 100% hit rate moves.
Not to mention Xcom.
Missing a 90% shot 3 times is so dam annoying.
In gen 1 that would be only Swift, Bide and self-targeting status moves because every 100% accuracy move can miss due to a bug (1/256). Fun fact: you would actually be able to beat the game with these 2 moves because Bide, in gen 1 only, bypasses Ghost’s immunity to Normal moves.
Hot take. FireRed > Red so I never need to play gen1 even if I am feeling nostalgic.
Mechanically, the remakes are objectively the better games and they also offer more gameplay-wise with new moves, more logical movesets, abilities, nature, actual EV system. The next big improvement was the physical/special split in Gen5.
I still play the gen1 games from time to time tho. They are hilarious.
It’s a big complaint for new players to Baldurs gate 3. People think a 95% chance to hit won’t fail but it does sometimes. Just the luck of the dice.
‘drop rates’ are the same thing. 10% chance doesn’t mean you will see it drop if you run a mission or defeat that boss ten times.
Or thinking that if you didn’t get the drop in 9 attempts, you’re practically guaranteed it in the next attempt.
Nope, still 1 in 10 chance.
In most simple written RNG calculations, past failures do not guarantee future success.
I believe some games will keep a tally of failures and award a successful loot after x failures to avoid frustrating players.
BG3 literally has setting for this.
Don’t use it though. The karmic dice system works for enemies too. So if you enable the system your rolls will fail less often but so will the enemy’s dice rolls. With karmic dice on I find the enemies crit me more. Especially on tactician mode.
Yeah Pity systems are quite common Also used in loot boxes to keep you buying with the terrible odds they have
I learned this very VERY well when I played Ragnarok Online in my teens.
There was a huge fight about this on The Battle for Wesnoth
Same thing happened with the iPhone shuffle. People complained it wasn’t “random enough” and would often end up calling members of the same family and/or household in a row. So they rewrote that algorithm too.
Wait, what?! There’s a call a random person feature? What a nightmare
Those were the best days of Apple. Steve Jobs literally titled the feature “Less Random”.
It’s not just about not wanting random, but randomness is actually very hard to create. Every random number is actual pseudo random
Some basic breakdowns of this concept:
https://slate.com/technology/2022/06/bridle-ways-of-being-excerpt-computer-randomness.html
The problem modern computers have with randomness is that it doesn’t make mathematical sense. You can’t program a computer to produce true randomness—wherein no element has any consistent, rule-based relationship to any other element—because then it wouldn’t be random. There would always be some underlying structure to the randomness, some mathematics of its generation, which would allow you to reverse-engineer and re-create it. Ergo: not random.
Kid friendly version:
Every random number is actual pseudo random
No, there are true random sources in a computer. Any outside input can be used to generate randomness. Mostly user input, but temperature fluctuations can work as well, if the sensor precision is high enough.
Also the argument is only correct on a technical level for PRNGs. Choose a 65535 sided dice and make the instructions a thousand steps long and you’ll have a pretty hard time to deduce the instructions from the generated numbers. Not to mention how long the list of numbers needs to be for the attacker to start guessing.
It is all based upon calculations with known numbers.
A computer can’t create a number out of nothing.
That is why Cloudflare uses lava lamps to generate random numbers for their cryptography. And even those aren’t fully random.
Modern cpus actually do have trng hardware built in. So yes, modern computers can create numbers out of nothing, because they have specialized hardware to do so
No, CloudFlare doesn’t use lava lamps to generate random numbers, that was a marketing stunt. Using a camera in a completely dark room is a better source of entropy than one pointed at lava lamps.
Also, nobody is saying that computers create a number out of nothing. The environment is a great source of entropy (temperature fluctuations, user inputs and so on) which are then expanded into a larger amount of entropy through CSPRNGs.
Using a camera in a completely dark room is a better source of entropy than one pointed at lava lamps.
Why is that? Naturally occurring or manufacturing-related impurities in the optical chip?
All digital cameras are imperfect - there is always a bit of noise, but usually it doesn’t come through since your scene is bright enough to make small amounts of noise imperceptible. In a completely dark room the camera still tries to get data from the photo sensor, but the noise (created by temperature fluctuations, imperfections in the chip itself and so on) is all you get. You may theoretically be able to predict the noise on short time scales, but it’s a chaotic system.
For the purpose of shuffling a playlist pseudo random is indistinguishable from truly random in all the ways that matter anyway.
This is an irrelevant distinction for any case where you aren’t worried about someone reverse engineering the algorithm and seed by logging output. Any half decent PRNG’s output will be statistically indistinguishable from true randomness.
Well, yea.
And since they removed my playlist radio, song discoverability has gone down so much. They just keep playing me the same damn music when I want related stuff.
For the life of me, I cannot figure out why they got rid of playlist radio. It makes Spotify so much worse and I can’t imagine it saves them any money
Probably some over complicated statistical model that will eek out cents more revenue per user.
It’s one of the worst changes they’ve made and it’s legitimately been making me take another look at all of their competitors.
Why does any corporation decide anything ever? It’s the money, Lebowski
For the same reason all the streaming sites decided to gimp their discoverability and just force you into whatever the algorithm thinks you need.
Some middle manager wanted to justify their bonus so they jumped in the current fad
I don’t use any streaming service, instead YouTube Vanced or NewPipe when I want to make a queue… But I absolutely hate that the algorithm chooses whatever it wants no matter what you play, and it’s just randomly chosen out of one of the few songs I played before. Also I sometimes play the one+ hour long compilations either to sleep (Nemo’s Dreamscapes) or fantasy music and YouTube decided that auto play will always choose one of those after playing anything (rock, metal, country, folk, name it, it doesn’t matter)
Wait until he finds out DJ X just plays the playlists from Spotify curators and he isn’t really keeping track of what you skipped.
I don’t know where you heard that but it’s not correct.
https://engineering.atspotify.com/2014/02/how-to-shuffle-songs/
I didn’t hear it anywhere, its personal experience. Spotify straight up isn’t shuffling any of my older liked music. Everytime I hit shuffle it always ends up localized around recently saved music and it never changes. Moreover, that’s a blogpost from a decade ago. Spotify changes stuff constantly. I seriously doubt that is the same algorithm they’re using now. ESPECIALLY considering they’ve introduced ‘Smart Shuffle’ which doesn’t only shuffle music you have saved but adds completely new music into the mix as well. That and Spotifys overall quality control has dropped drastically in the past decade.
While I can agree with others that it might not be that strict and they may get more variety, from personal experience, my 700+ song playlist tends to focus on a handful of songs. Blue - Eiffel 65 gets a CRAZY amount of playtime for some reason. It’s a running joke with my carpool that it’s my car’s theme song. I can sometimes skip it 2-3 times on my drive to work. And some songs seem to never come up. I’ve been having some better luck with the AI DJ lately, although it will throw in an entire chunk of songs I don’t even care about as basically ads.
That song is probably cheap for them to play
I’ve got thousands of liked songs and I get a nice variety of old and recent stuff. From last month to as old as 8 years ago. Sometimes I get too many songs from the same band, sure, but when you’ve got 50-100 songs from the same band liked, that’s bound to happen
Yeah I’ve got like 8000 and there have been times when it feels like it’s focusing on newly added stuff, but older stuff definitely does pop up and I feel like it’s a pretty good mix. 90% of the time I’ve just got my whole library on shuffle unless I’m actively looking for new stuff to add.
Personal experience aka anecdotal evidence aka fucking useless
Especially when they state it like it’s fact in the title and then can’t back it up at all.
Oh really? Have you ever eaten a peanut butter and jelly sandwich?
Nope, what’s your point?
I use Virtual Shuffle to force Spotify to play truly random tracks https://shuffle.virock.org
Yeah okay, so they use the Fisher Yates algorithm. Then how the fuck in my playlist of 1000 songs am I CONSTANTLY getting hit with multiple songs by the same artist? Once in a while, okay, that’s statistically likely, but not all the time, and not every time.
how many songs by the same artist do you have in that playlist? Your comments here don’t make sense. Repeating the same artist doesn’t make something not random and definitely doesn’t make it “the most 100 recently saved songs”. In fact repeating the same artist makes it quite likely that it is random as randomness is frequently misinterpreted as non-random if you would have read the article I linked you would understand that.
Interesting read
The shuffle button does it’s thing every time you activate it. I believe, anecdotally, that it uses whatever song is playing or selected as a seed to build the random queue.
Try this, select a song and press Shuffle. When it gets to a song you don’t want to hear, skip to a song you want to hear and toggle Shuffle off and on.
Good to know, but it would be even better if it just shuffled the playlist you are listening to instead amaright?
But you can make it actually random by doing this:
- Tap your profile picture
- Tap settings
- Tap Playback
- Toggle the “Automix” switch to off
Now you have truly random shuffle
I’m pretty sure this is wrong. From what I understand, automix just blends together the ending and beginning of the two songs. For example if there’s silence at the end of one song, or ‘compatible’ music, the ‘ai’ will try to smooth out the transition like a dj would while making a mix.
Edit: yup, that’s what it is https://support.spotify.com/ca-en/article/tracks-transitions/
Turn off* cross play.
That will give more randomness, but I still get a bunch of songs on repeat in a 3K song playlist. Annoying.I followed this life hack article and it worked for me: https://lifehacker.com/the-reason-spotify-shuffles-aren-t-really-random-and-h-1849756947
Often completely botching existing transitions on, say, two consecutive songs on an album that already flow into one another.
Well it’s some setting. I recently made the switch and now I’m hearing random songs I haven’t heard in years. It’s actually sampling my whole collection as far as I can tell.
I doubt it. True random shuffle play is rare, because humans don’t understand the chaos of true random generation, we see patterns in it and assume it’s not random.
A truly random shuffle can play the same song twice. A truly random shuffle can play multiple songs from the same artist in a row. In the fullness of time all of these will happen with a true random shuffle.
Nothing does that these days. Nearly everything “random” is algorithmically engineered to be less random so it feels more random to humans.
You can definitely make a truly random shuffle that doesn’t play the same song twice within a gap of ten songs.
You just make a list of all the song IDs that aren’t in the last ten played, and grab an item at a random index.
Randomness is a texture not a shape. Dice are random despite only being able to present six outcomes.
But then it is not truly random, that is the point.
The same song 10 times in a row also has a chance of happening in a truly random list.
Random pick without replacement is no less random than random pick with replacement. You just have a continuously smaller pool to pick from.
It’s absolutely possible to have better randomness/shuffle without repeating songs so often.
there exist other random distributions than the uniform distribution
You’re thinking of a random sequence. We’re talking about shuffling according to a random seed.
now I wish for a more nuanced configed shuffle because I totally love having a few songs from the same artist back to back but not the same one.
A truly random shuffle can play the same song twice. A truly random shuffle can play multiple songs from the same artist in a row. In the fullness of time all of these will happen with a true random shuffle. Nothing does that these days.
Tidal does and it’s annoying af.
A truly random walk through a playlist might choose the same song twice in a row. A truly random shuffle would only have each entry appear once and you’d have to play past the end of the shuffle to hear a song repeat.
you my sir are godsend
Shuffle uses a limited list that rarely gets updated, but not just the 100 most recent ones. You can force it refresh by turning off shuffle and force closing the app.
Hit the shuffle button twice to turn on Smart Shuffle. It will throw some brand new songs into the playlist shuffle.
deleted by creator
I’m on android. I just confirmed you can’t do it with an album but instead with playlists. Add a song to a playlist, hit shuffle, then hit shuffle again and a little star should appear in the shuffle icon.
Only works on self created Playlists, not on playlists created from spotify like “Discover weekly”. The first time, you get a little popup which explains it.
Can confirm works on android, no clue about Apple tho.
It is shit like this that made me get back into buying my own music.
On the desktop spicetify solves this with the shuffle+ plugin. On mobile you can browse to a file in a playlist and start from there instead of shuffling the playlist. Turning off Automix also helps.
Nah, it’s payola.
I’ve been listening to music from the 60s almost exclusively on spotify this year. The algorithm is hell bent on me listening to the band “The Smoke”. I really don’t think a one hit wonder band that broke up in 1976 would be paying to have their music pushed so hard like that.
It’s probably for the best
I specifically use shuffle on my favourite list when I get tired of the music rut I’m in and want to listen to the shit I listened to 5-10 years ago
there was that meme about adding everything to one playlist a few days back, that’s mel, lol, I don’t even know thr number of songs on the Playlist, it’s 112 hours long at the moment.
Technically right. Computer can’t do actually random calculations, instead researchers have made pseudo random functions, which similarly behaves as a truly random function would.
Spotify is still the worst in that regard. It even loves to constantly repeat the same ~5 songs or so, at least before I quit the damn thing. Winamp has done this better a couple decades ago already, among many other features.
That’s true. It also has nothing to do with the topic: only shuffling recent songs (allegedly). The title has more info than the meme image.
Why the fuck is this downvoted??? They are absolutely right
A link for the uneducated and ignorant: https://slate.com/technology/2022/06/bridle-ways-of-being-excerpt-computer-randomness.html
They probably didn’t like my nerdy joke. I see the argument that it’s kinda not the topic, but we are talking about randomness. It sounded fun in my head.
Because it’s not the reason. Pseudorandom number generation is statistically identical to random.
deleted by creator
deleted by creator
Yeah because that’s reasonable… give me a fuck break
How many times you gonna sit there thumbing down tracks when listening to random? What if you actually like the song?
This ain’t a solution
Spotify has a thumbs down button?
