🧵 Untitled Thread
Anonymous at Fri, 13 Sep 2024 14:16:34 UTC No. 16378558
Why is Law of large numbers a Law? Where's the proof? What is a sufficiently large number? If I consider the results of a large number the single outcome of a bigger scale trial, wouldn't that imply it would give me erratic results and contradict itself?
Anonymous at Fri, 13 Sep 2024 14:43:08 UTC No. 16378597
>>16378593
Sorry I don't speak greek.
Anonymous at Fri, 13 Sep 2024 18:35:34 UTC No. 16378946
>>16378597
tl;dr
Dividing by 1/n shrinks the variance faster than adding n stuff, so as n gets bigger it becomes closer and closer to 0 variance
something with 0 variance is just a non random number
you can read the proof to construct the desired value of n so that the variance is tightly constrained and the random fluctuations are only epsilon big
Anonymous at Fri, 13 Sep 2024 20:27:10 UTC No. 16379110
>>16378946
Oddly fascinating lol
Anonymous at Fri, 13 Sep 2024 20:29:56 UTC No. 16379115
>>16378946
dividing by 1/n would literally explode the variance to infinity wtf are you smoking
Anonymous at Fri, 13 Sep 2024 22:18:46 UTC No. 16379229
>>16379115
NTA but you should do some algebra to check it out it might blow your mind
Anonymous at Fri, 13 Sep 2024 22:39:19 UTC No. 16379252
>>16379229
How would dividing n by 1/n shrink the variance? Think about and it might blow your mind