Hi /sci/
There's something about chance that I just don't get.
Say there's a system with 1/4 chance of printing A instead of B. If it prints four B's in a row, won't it be more likely that there's an A coming next? I mean the chances still are 1/4 in theory but it feels as if the chance of A increases with each printed B.
Is that just perception's fault?
>>7934873
each event is independent.
just because your specific string only has a chance of 81/256 to occur doesn't mean it changes the probability of the system in general. It still has a 1/4 chance to print A and a 3/4 chance to print B at each printing step.
>>7934895
But what about luck?
Luck defined as the curve in probability according to the one who inputs the beguining of a sequence. As no matter is lost or destroyed, the universe exists on a very specific but very very very complex numerical pattern. Assuming that holds true, the pattern must be upheld at all times in the universe, and when a high probability outcome occurs anywhere in the universe, there must be a low prob outcome elsewhere regardless of form.
It is the ULTIMATE game of getting dubs, trips, quads, pents, hexs, etc but IRL.
So in a wager situation, if you watch all the events of the world closely and recognize their numerical and probabilational value, if you time it right you can win very low probability outcomes everytime.
That begs the question, if I can cheat the system then I have 100%, but theres still a probability of me getting caught, is probability an arbitrary concept?
>>7934913
Nope, the events are independent and "luck" does not exist.
>>7934873
> won't it be more likely that there's an A coming next
Only if the relationship between the random variables dictates that.
What you're talking about here, is conditional probability. The equivalent question is
> given four B's were just printed, what is the probability the next print is an A?
Depending on the physical phenomenon, the probability could be greater, lower, or it might not change at all. If it doesn't change, then you have independent events.
At the root of your question though, is a fundamental misunderstanding of "randomness".
>"If the probability of thing is 1/6, then I just try 6 times and it'll have to happen!"
That's not how randomness works, since that basically describes a deterministic system (i.e. "if I do thing a certain way, I know exactly what will happen with full certainty").
> Luck defined as the curve in probability according to the one who inputs the beguining of a sequence
You might be thinking of the likelihood ratio, which kinda describes the change in the probability of an event happening vs not happening.
> the pattern must be upheld at all times in the universe
... but it actually sounds like you're just spouting pseudo-philosophical BS at this point.
>>7934913
Stop posting immediately
>>7934913
bad troll
>>7934895
What I understand by 1/4 chance is from a big sample, 25% were A's, all others are B's. Now, could it happen that out of 1000 prints, not a single one turns out to be an A? It could apparently, as 25% does not imply that it must have happened after certain number of prints. Though if that's the case then I don't understand what chance is.
>>7935007
>> given four B's were just printed, what is the probability the next print is an A?
>conditional probability
this sounds like the kind of term I'm looking for.
What I'm trying to ask then is, at any specific point, would the probability of the next A increase/decrease depending on how many B's came before it?
>>7935063
>>7935067
OP here. I'm not >>7934913
>>7935119
I think you are correct in your statement that you have a poor understanding of 'chance'. The reason to this can best be explained by answering your question:
>"at any specific point, would the probability of the next A increase/decrease depending on how many B's came before it?"
The answer to this is *without question* 'no'.
Lets look at an easier case, maybe it will help you understand. Lets say I'm flipping a coin. It has ~50% chance to land heads up or tails up either flip. So if I flip it 100 times, I should expect around 50 heads and around 50 tails. BUT it is completely possible for it to land heads up every single time. If I then flip this coin again, for a 101st time, there is still a 50% chance it will be heads, and a 50% chance it will be tails.
When youre dealing with true randomness or chance, there is no such thing as patterns. There is no such thing as numbers 'being expected' either because they've come up a lot or because they've come up not at all. Every event is its own instance, where you must go back to the original odds. If you need to change your current understanding of chance, I recommend doing so, because I promise you this is how it works.
I'm this poster:
>>7935487
Continuing on:
The only situation were this would not be the case is one were you change the semantics of the question. For example, if instead of saying "There's a 25% chance of printing A instead of B" you said "25% of the time it will print A instead of B" then you can begin to draw conclusions on the likelihood of the next print to be A or B.
If we know 'for sure' that 25% of the time it *has to* be A, then there are things we can do. For an easy example, if there were 75/100 questions printed so far and all of them were B, we can say with 100% certainty that every remaining answer will be A.
Or, if we print 24/100 and they are all A so far, we can say there is a greater than 98% the next question will be B. So it *might not* be B next, but its considerably more likely it will be B.
Where as in the original question (there is a 25% chance of printing A instead of B) if 24/100 were A so far, the next question *still* has only a 75% chance of being B. This remains unchanged.
>>7935119
Let's do a thinking problem.
Grab a coin. Flip it until you get 4 heads in a row.
Stop.
What are the chances the next flip is Tails?
Now, instead of flipping, put the coin down.
Come back in 3 days.
Pick up the coin. What are the chances that the flip is Tails?
Now, find a random coin anywhere. What are the chances that you will flip heads and tails? Do you know what the past 4 coin flips were? The past 40? 4000? Does it matter?
Now, think of the first coin you picked up to flip. What were the head/tails chance when you started flipping it? What were the last 10 coin flips, however many weeks/months ago before you picked it up to flip it. Did it matter? Does luck get "stored" for certain amount of time before you lose it?
No.
>>7935510
Not OP, but I have a question, how does one go about scientifically actually determining probability of something like a coin or like dice? Say I have a coin that after a thousand flips comes up 80% tails and 20% heads, basic understanding tells me that coins are said to have a 50/50 chance no matter what, where does this actually come from? (in my 80/20 case should I start looking for coin deformations or just assume that it wasn't 50/50 this time but that in the future it will be because it is supposed to be)
Is it possible that a side with a head engraving on it that weights more might make the odds of a specific coin actually be 50.002/49.998 or something? Where do the proclamations that it is exactly 50/50 come from? some asshole pulling out of the air because it seems "close nuff"?
If I knew that it was actually 49.5/50.5% in actuality wouldn't that change my perspective on which side I should call in a given coin toss?
>>7935515
Poot understanding of luck, is not a karma system or a bank account, it's just the status of always encountering the rarer outcomes of things with equal chance.
Like how a surfer doesn't just stay up because he's surfed before and has credit, he's conditioning himself along the wave every second of the way. He falls right into the grooves the board already fits in.
>>7935576
>Not OP, but I have a question, how does one go about scientifically actually determining probability of something like a coin or like dice? Say I have a coin that after a thousand flips comes up 80% tails and 20% heads, basic understanding tells me that coins are said to have a 50/50 chance no matter what, where does this actually come from?
Experimentally determining probability from empirical data involves updating a hypothesized prior probability. Read
https://en.wikipedia.org/wiki/Bayesian_probability
But keep in mind that the 50/50 coin is simply an idealization for the purposes of thought experiments. Obviously a coin in real life can have features that change that.
>>7935577
Put down the bong, retard.
>>7934913
>>7935576
Interestingly enough:
http://www.smithsonianmag.com/science-nature/gamblers-take-note-the-odds-in-a-coin-flip-arent-quite-5050-145465423/?no-ist
Its actually 51/49 not 50/50. The bias is towards whichever side started up when you flipped. Or alternatively, if you spin it, its considerably more likely to land tails up due to the weight difference.
>>7935487
>There is no such thing as numbers 'being expected' either because they've come up a lot or because they've come up not at all. Every event is its own instance, where you must go back to the original odds.
This really helped man, thanks. I think I got it mixed up with how chance is determined. I see you respond to that here >>7935618 I'll give the link a go-- it does look a bit daunting to me tho, I've never taken any formal class in a related matter yet.
>>7935515
I enjoyed this thinking problem mate. Made me feel a bit stupid, which is a good sign that it clicked. Cheers!
Thank y'all
>>7935576
The probability of a coin or dice has to do with how many possible outcomes there are. A coin has two sides and the coin is symmetrical, so there really shouldn't be any reason one side will come up more than another. so the chances of getting one side is 1 in 2, or 50%. A die has 6 sides so the chances of getting one side is 1 in 6, or like 17%.
>>7934873
It is kind of fucked with Computers, they are programmed to give you a chance but in real life it's random
>mfw 1-2-3-4-5-6 has exactly the same probability of being drawn in a lottery as any other permutation
>>7934873
well if a sheet is printed n times and has a 1/4 probability that it will print A instead of B, then the probability of not getting an A decreases as n beomes bigger. each event is still a 1/4 probability but when looked at as a series getting at least one A becomes more probable
>>7938132
But, 123456 is the most commonly chosen series of numbers so you'd be sharing your prize with tens of thousands of other people
>>7935119
Independent probabilities are independent for a reason. What you are thinking of is conditional or successive probabilities, i.e. the chances of you drawing an ace from a deck of cards 3 times in a row is (4/52)^3
The world is indistinguishable from random without a narrative to delineate objects and relationships, and that all depends on a purpose.
The question isn't whether it seems like the probabilities should be accumulative; the question is how you looked at the world in the first place makes how you see the world.
So no. It is not surprising that you think the probability should change. The Meta-narrative about how you think would suggest that a different narrative is at work: You are simply doubting that the probability is indeed 1/4 after 4 B's in a row.
The ability of your brain to detect patterns through the narrative is not static; it is dynamic in that you never "know" something, but you are always "learning" something.
>>7938151
Can anyone back this up? It seems like a nicer version of what I tried to refer to originally .
>>7937933
What about the chance in the example. 1/4 for A but only 2 possible outcomes?
The chance of rolling 6 from a fair dice tends to 1/6 as more times you roll. Would you have to do the similar to come up with '1/4' for A's chance? Print a fair amount and see if it's approaching some value?
>>7935487
^I'm this guy, back here to confirm that:
>>7938151
^This post is accurate, but maybe not quite in the way you think it is. The poster did word their explanation very well, but I'd like to expand on it a little.
>when looked at as a series getting at least one A becomes more probable.
This is *only* true when looked at as a whole series. Before you have printed any letters, that is the only time this holds true. Once you start printing you need to change how the odds work accordingly.
> each event is still a 1/4 probability
This is the key. Once you start printing, each individual event ALWAYS has a 1/4 probability. If you are 1,000,000,000 prints in, and there hasn't been a single A yet, this has A *absolutely no effect* on the outcome of the next print. The next print still has a 1/4 chance to be A and a 3/4 chance to be B.
>>7939706
We throw a coin that comes up heads with probability 1/4. If so, we print A.
New sheet. We throw a coin that comes up heads with probability 1/4. If so, we print A.
We do it n times. What is the probability non of our printed sheets had A on them?
(3/4)^n. What is the probability at least one had A on them? 1- (3/4)^n. This tends to 1 as n increases. What you really need to understand, and what you might want to read up on, is what independence means and why, when we flip coins, it is an appropriate formalisation of the experiment.
You might also want to read this:
https://en.wikipedia.org/wiki/Geometric_distribution
>The chance of rolling 6 from a fair dice tends to 1/6 as more times you roll.
No, for any roll, the chance is 1/6. What tends to 1/6 as you roll is the number of times you got a six divided by the number of rolls. That is implied by the fact that each roll has a chance of 1/6 to be a 6 (this is the law of large numbers), but does not imply it.
>Would you have to do the similar to come up with '1/4' for A's chance? Print a fair amount and see if it's approaching some value?
Again, if we think of an event that has a probability of 1/4 of occuring, we can calculate any probability we want. If we just observe As and Bs on a paper and we assume each occur with a fixed but unknown probability, we can estimate this probability in the manner you describe. But this is statistics, not probability.
In probability, we just say what can happen and then we state what the probability is we assign to every possible outcome. At least if outcomes are discrete. This is what OP did. He said A can occur or B can occur, with probabilities 1/4 and 3/4. He also talked about a setting where a natural assumption is that events are independent. From that, we can calculate the probability of any patterns or count of As or Bs, of interest.
>>7934873
gamblers fallacy look it up
