/sci/ - Science & Math » Thread #13574879

4KiB, 199x253, download (1).jpg

View Same Google iqdb SauceNAO

Anonymous Sun 29 Aug 02:44:24 2021 No.13574879 View Reply Original Report

Quoted By: >>13574894 >>13574905 >>13574907 >>13574958 >>13575004 >>13575041 >>13575313 >>13575425 >>13575443 >>13575486 >>13575733 >>13575744 >>13575770

how many times does something need to happen for you to be certain of the probability?

example if you flip a coin 100 times and you get exactly 40 heads 60 tails, should you just decide that theres a 40% chance to get heads?

Anonymous

Anonymous Sun 29 Aug 2021 02:52:32 No.13574894 Report

Quoted By:

>>13574879
>how many times does something need to happen for you to be certain of the probability?
i am a brainlet retard but here's what i think.

the probability of a coin flip isn't based off what you see. a coin flip is 50/50 is because there are two possible outcomes, and each outcome is equally likely.

so your example of getting 40 heads out of 100 flips, that doesn't mean 40% probability for heads, that's just getting heads 40% of the time. probability is not the same as what you get. does that make sense

Anonymous

Anonymous Sun 29 Aug 2021 02:56:17 No.13574905 Report

Quoted By:

>>13574879
https://en.wikipedia.org/wiki/Law_of_large_numbers

Anonymous

Anonymous Sun 29 Aug 2021 02:56:41 No.13574907 Report

Quoted By:

>>13574879
>how many times does something need to happen for you to be certain of the probability?
Infinite times, bro. You don't know if that coin will forever land on heads on all coin flips only after the millionth flip.

Anonymous

Anonymous Sun 29 Aug 2021 03:03:50 No.13574931 Report

Quoted By: >>13574952

i think there should be several methods to figure this out, but theres no definitive answer.

the first thing you need to do is have a define at what point should you accept an answer

like you expect that the coin should have 1/2 chance of getting either head or tails, but if something thats rarer than 1/1000 chance of happening happened, say you got 100 heads out of 100 flips, then you just decide its not actually 1/2 chance of either head or tails

Anonymous

Anonymous Sun 29 Aug 2021 03:11:44 No.13574952 Report

Quoted By:

>>13574931
>but theres no definitive answer.
there is though. it's 50/50.

>but if something thats rarer than 1/1000 chance of happening happened, say you got 100 heads out of 100 flips, then you just decide its not actually 1/2 chance of either head or tails
no. that's not what "chance" means
just because something unlikely happened doesn't mean it was likely to happen

you have a "right" idea but you're using the words "chance" and "probability" wrong. the chance and probability doesnt change

Anonymous

Anonymous Sun 29 Aug 2021 03:14:45 No.13574958 Report

Quoted By:

>>13574879
probability is a predictive measure anon, not a historical one, which is statistics and data. As far as probability goes, it is always 50/50 for a coin toss even though it is accurate to say that historically speaking that specific coin has favored one over the other

Anonymous

Anonymous Sun 29 Aug 2021 03:28:57 No.13574998 Report

Quoted By:

simple question but way too difficult for some anime messageboard neckbeards

Anonymous

Anonymous Sun 29 Aug 2021 03:31:33 No.13575004 Report

Quoted By:

>>13574879
>certain
You can't reach absolute certainty but you can get a probability density function that approximates it.

Anonymous

Anonymous Sun 29 Aug 2021 03:38:07 No.13575028 Report

Quoted By: >>13575456

A coin is a great example of how confusing probability is. What are the chances of getting 100 heads? Very small. What are the chances of getting tails after you just got 100 heads? Still 50%. Every individual is 50% which is precisely why sometimes you get heads maybe ten times in a row.

Anonymous

Anonymous Sun 29 Aug 2021 03:43:00 No.13575041 Report

Quoted By:

>>13574879
Confused question, take the bayespill

Anonymous

Anonymous Sun 29 Aug 2021 05:26:01 No.13575313 Report

Quoted By:

>>13574879
Certainty of probability doesn't exist because there's the probability of me flipping the coin infinite times and getting heads all the time, it would only be a certain a "50/50" if heads and tails happened at the same time

Anonymous

Anonymous Sun 29 Aug 2021 06:11:45 No.13575425 Report

Quoted By:

>>13574879
You can’t be *certain*

Anonymous

Anonymous Sun 29 Aug 2021 06:22:16 No.13575443 Report

Quoted By: >>13575475

>>13574879
You cant prove with 100% certainty any outcome based on finite attempts
Thats why science uses relative probability, and the concept of sigma to evidence their claims

Anonymous

Anonymous Sun 29 Aug 2021 06:29:14 No.13575456 Report

Quoted By:

>>13575028
>A coin
you should specify fair coin to be honest

Anonymous

Anonymous Sun 29 Aug 2021 06:35:50 No.13575475 Report

Quoted By: >>13575477

>>13575443
whats sigma

Anonymous

Anonymous Sun 29 Aug 2021 06:36:17 No.13575477 Report

Quoted By:

>>13575475
Sigma balls

Anonymous

Anonymous Sun 29 Aug 2021 06:39:04 No.13575486 Report

Quoted By:

>>13574879
You can't be 100% certain. But based on your example you could calculate a certainty that's very close to 100%, that the values you want to know are within a certain range.

I'm not going to actually do the math.

But based on your example the answer might be: after flipping 100 coins and getting 60 heads, there is a 99% certainty that the coin lands on heads 59% to 61% of the time.

The actual numbers will be slightly different, but that's the form the answer will take.

Anonymous

Anonymous Sun 29 Aug 2021 08:19:36 No.13575733 Report

Quoted By:

>>13574879
This is kind of what the entire field of statistics exists to do
https://en.m.wikipedia.org/wiki/Statistical_hypothesis_testing

There's no simple answer to your question. But in general, you start with an assumed distribution and work from there to check if the data fits it, and how to adjust your distribution if it doesn't. A lot of these methods are still contentuous.

Anonymous

Anonymous Sun 29 Aug 2021 08:23:09 No.13575744 Report

Quoted By:

>>13574879
an asymptotic amount of times

Anonymous

Anonymous Sun 29 Aug 2021 08:32:34 No.13575770 Report

Quoted By:

>>13574879
>how many times does something need to happen for you to be certain of the probability?
It's never certain.

>if you flip a coin 100 times and you get exactly 40 heads 60 tails, should you just decide that theres a 40% chance to get heads?
I don't know. What's the cost of flipping, and what's the cost of getting it wrong?

Anonymous

Anonymous Sun 29 Aug 2021 08:58:53 No.13575848 Report

Quoted By: >>13575935

OP here tried to look this up, found a bunch of confusing literature but no examples. Is this supposed to be a difficult problem or is there just a formula to figure it out

I get that it can’t be “certain” now, but how do you for example determine that it’s a 40% chance with 99% accuracy? How many trials for that!

Anonymous

Anonymous Sun 29 Aug 2021 09:58:21 No.13575935 Report

Quoted By:

>>13575848
>how do you for example determine that it’s a 40% chance with 99% accuracy?
Well, suppose you are wrong. Suppose it's actually 50-50. What are the odds of getting a 40-60 from a fair coin after 100 attempts?

Capcode	All Only User Posts Only Moderator Posts Only Admin Posts Only Developer Posts
Show Posts	All Only With Images Only Without Images
Deleted Posts	All Only Deleted Posts Only Non-Deleted Posts
Ghost Posts	All Only Ghost Posts Only Non-Ghost Posts
Post Type	All Only Sticky Threads Only Opening Posts Only Reply Posts
Results	All Grouped By Threads
Order	Latest Posts First Oldest Posts First

Your latest searches