Message from @MOS MAIORVM (AKA Sola)
Discord ID: 311379037337550848
It matters to a point
It matters period
Depends on what you study
No. It doesnt matter period.
Yes, we only have so many resources to expend on it
Per the 100k vs. 10 mil example
But it's important to get a large sample size
Maybe. It certainly isnt some inviolable principle. Maybe it is useful.
10 mil will still give you a more accurate result than 100k
It doesnt make it MORE TRUE
10 mil vs. 100k in coinflips
It's not about being "more true" but more accurate
Will not give you a more accurate result
it will
More accurate what though
Than what you'd likely otherwise get just doing 100k
in other words
Hang on
sure thing
100k fully establishes a certain premise
and that premise is valid
How can you get more accurate than 50 50
why do you assume you'll get a perfect 50/50 at 100k?
Will it become 50.0n1 on one side?
You likely won't
but you'll get close
You will.
and you'll get closer at 10 mil
Why stop at 10 mil
You believe someone will always flip exactly 50k of their 100k flips on heads every time they attempt it?
Go for broke
you can stop at whenever you want
10 mil study under the same circumstances will be appreciated more than the 100k
and there's valid reason for it
No. But you can nevertheless extract the trend at 100k. 10 mil helps you not one bit.
Not even a little bit
Well it gets you a more accurate result
Its ridiculous you think so
Whether you think it's accurate enough is subjective
This is exactly what I meant