Wheel of Fortune and Space Joker explained (No it's not true 1 in 4)

[Mcholypotato]

The game uses a seeded random approach, meaning the outcomes of cards like Wheel of Fortune and Space Joker are predetermined when the run is generated. The "1 in 4 chance" for Wheel of Fortune, for example, is already calculated and fixed based on the seed that governs the run. This is different from what’s called dynamic RNG, where a probability check would be performed every time the card is used during gameplay.

To help visualize this, think of a game like Minecraft, where entering a seed generates a specific world. The locations of resources like mining nodes are predetermined by that seed. No matter how many times you reload or revisit, the resource locations will stay the same unless you use a different seed. Similarly, in Balatro, the outcomes for cards like Wheel of Fortune are "baked in" by the run's seed. This means you could have runs where the card procs nearly every time and other runs where it barely procs at all, entirely depending on the seed.

Seeded randomness is great for consistency and reproducibility (which is why you see seed numbers at the end of Ante 8), but it can feel less "random" to players because the outcomes are predetermined and don’t change in response to real-time actions. This can result in streaks of success or failure that feel frustrating, even though they’re just part of how the seed logic works.

TL;DR: The outcomes for Wheel of Fortune and Space Joker aren’t truly random because they’re determined by the run’s seed.

[ulzgoroth]

Seeded randomness cannot feel less random to players who aren't save-scumming or something (which is difficult at best in Balatro). Because you have no way to perceive it.

"True" randomness also doesn't change in response to real-time actions. They're random. True randomness frequently results in streaks that feel frustrating or shocking, because the dice do not care about your feelings, and the gambler's fallacy is a fallacy.

[snakeskip]

Every PRNG is seeded. Randomness in games is almost always seeded just like it is in Balatro.

Some games just don't allow entering seeds, don't show seeds to the player and don't save seeds (or states of PRNGs). Balatro does.

But that does not change probabilities. 1/4 is 1/4. There is no measurable difference from the player's perspective. Other than save scumming not working since seeds are saved.

Besides, Balatro actually does probability checks as "dynamically" as any other game.

The only special thing Balatro has to do is to keeps RNGs for different elements separated. That is needed to keep seeded runs "stable". Seeded runs would be pointless if playing a Lucky Card would change the contents of the next shop etc. Players playing the same seed would not see the same shop contents, bosses, shuffles etc after the first Ante or so.

[Goblin]

Everything you said is correct, except the part where you claim this makes it "not truly random."
The fact that the randomness comes from a seed doesn't change the fact that it's a 1 in 4 chance. If I roll a standard die, cover the result, look at it, and tell you you have a 1 in 6 chance of it being a six, that statement is still correct. Even if I've already seen it's a two. Doesn't matter that the result has already been determined, the odds are still the same. The die roll is still just as random.

[Dan]

It seems like this is based on an assumption that the seed will be applied independently and in isolation to each individual card. Have you or anyone else actually tested and confirmed this assumption? - Because that is *not* typically how anyone would implement seeding in a game. Everything would be drawing from the same seeded sequence, so every time your deck was shuffled or any other card tested probability, it would alter the results of every other random effect in the game.

Also, you seem to be working on the assumption that there's an alternative to seeded random number generation. That literally *is* random number generation. The only difference in a "non-seeded" situation is a game/program simply sources a seed from your system clock milliseconds, or some other unpredictable source. Computers can't do chaos, they're deterministic logic processors. The good thing is, unless you're a doctorate-level mathematician, you couldn't possibly tell the difference. All standard PRNG algorithms in modern programming languages produce such high quality randomness that you'd need some fairly deep analytics to spot any kind of underlying pattern over billions of iterations. That doesn't generally have any bearing on anything other than scientific computing applications.

[opie]

this seems like more a concern of philosophy than of gameplay

[Twelvefield]

I agree, I don't understand why I should care about this. I appreciate the good, thoughtful writing, but I don't see to what end it serves.

Cards, digital or analog, are cards. The skill in cardplay comes either from mitigating possibilities or from cheating them. The entertainment in cardplay doesn't often come from discussions of computer code.

[ulzgoroth]

I assume the point here is implying that this explains why people get mad about Balatro randomness.

For which purpose it's wrong, because Balatro randomness is almost certainly not humanly distinguishable from the randomness you'd get from real cards or dice. The problem is the people, not the RNG.

[Kodero]

Its a fascinating topic, especially if you look at it on a deeper level. Everything in the universe is the result of cause and effect, and every effect will start a new cause. It is quite hard to understand this beyond the surface level of which we tend to focus. But if you can recognize this phenomen, cause and effect, you will also understand that nothing that happens is random.

Perhaps it is when we decide that we cannot longer see or understand the causes and effect we are content to call it random.

[ulzgoroth]

Originally posted by Kodero:

Its a fascinating topic, especially if you look at it on a deeper level. Everything in the universe is the result of cause and effect, and every effect will start a new cause. It is quite hard to understand this beyond the surface level of which we tend to focus. But if you can recognize this phenomen, cause and effect, you will also understand that nothing that happens is random.

Perhaps it is when we decide that we cannot longer see or understand the causes and effect we are content to call it random.

There's some reasons to think that's not actually true of fundamental physics. There's stuff that lines up well with the universe genuinely playing metaphorical dice.

A little less ludicrously digressive, the psuedorandom numbers used in games have cause and effect that are, ultimately, quite easy to understand. However, understanding doesn't break the effect - a good pRNG gives sequences of numbers that don't have patterns you can use except by knowing the key and following the algorithm.

[Ronald Brain]

TL;DR: The outcomes for Wheel of Fortune and Space Joker aren’t truly random because they’re determined by the run’s seed.

But consider that each unseeded run we start is randomized, isn't it the same? You're still rolling into a random seed which in turn rolls WoF / Space Joker's probability, and it is still 1 in 4 at the end of the day.

[pacman]

> Seeded randomness cannot feel less random to players who aren't save-scumming or something (which is difficult at best in Balatro). Because you have no way to perceive it.

> "True" randomness also doesn't change in response to real-time actions. They're random. True randomness frequently results in streaks that feel frustrating or shocking, because the dice do not care about your feelings, and the gambler's fallacy is a fallacy.


Research done by the Princeton Engineering Anomalies Research Program found that some people were able to raise or lower the statistical mean of a quantum random number generator by concentrating intently on it. So if you aren't getting enough procs you probably aren't believing hard enough.





(These studies are highly controversial and there have been issues with reproducing the results, but I choose to believe it because a world where you can shoot mind bullets is more fun than a world where you can't).

[CMDR Shven]

There's no practical difference between what you're talking about and a non-seed determined random outcome at every instance of using one of these effects. At the end of the day the cards will still succeed 25% of the time over the long term. The cards still do what they say they do.

[ulzgoroth]

Originally posted by CMDR Shven:

There's no practical difference between what you're talking about and a non-seed determined random outcome at every instance of using one of these effects. At the end of the day the cards will still succeed 25% of the time over the long term. The cards still do what they say they do.

Also, and this part is important, they pattern of success is going to actually statistically look like independent random trials. Sequential 'nope', 'nope', 'nope', 'success', repeat... would have the right average over the long term, but obviously it wouldn't be random.

[AnonTwo]

This is irrelevant unless you run the same seed more than once. The RNG is still rolled at some point.

[Adam Heron]

I can't believe after all this time we've found a new way to misunderstand probability. Nicely done.