The law of large numbers hummel

ok...that´s a strange answer to the question of why you don´t make use of this law, or did I miss something?

The law of large numbers in a nut shell means that no matter how many trials I use it wont matter because that can be just a small percentage when compared to something like infinity.

Say you take a coin and you flip it not just a few million times but a few billion, no a few trillion times. Well there is nothing stopping the results to end up being 75% heads and 25% tails because when compared to infinity even a few hundred trillion can be considered a small trial.

In other words you can't reliably determine the probability of something just by running trials. If you've ever played a game with a rare item drop you know how gay it can be.. a 1in100 drop might not trigger until the 275+ time or more. Or you might get it on the 1st kill.

put another way, the only way you'll ever truly get a 50% heads and 50% tails result 100% of the time is if you flip the coin an infinite number of times. There is no large number when comparing to infinity.

Then find a counterexample (i.e. with a minimal amount of balls) that the quote is wrong.

Trust me I've looked for it superku I'm having no luck. For some reason the only things I can find are wikiapedia entries which the math entries are always hard to read for me. I don't understand how they present the formulas.

Seams to be the opposite to what Wikipedia says:

Of course there is a chance to get 75% heads and 25% even with a few million trys but the probability converges to zero.
Also you have to compare the amount of trys to the amount of possible results you can get for one try. Not to infinity.

In less than 10 minutes of coding you could get reference results which would tell you what solution might be the right (if the random-func you use is good).

yea that's part of it but it should also go on to explain that it wont really matter how large a finite number is.



you make a perfectly valid point in that case.

I'd still would like to know with no doubt how to do this mathematically for my own person growth/education.

that's an easy question. just build a probiablity tree and add the branches that yield the wanted experimental outcome. shouldn't take you more than 5 minutes.