Quote Originally Posted by Pythonfriend View Post
thought experiment:


You get a black, intransparent bag, and are being told that there is a 50% chance that two white marbles are inside and a 50% chance that one white and one black marble is inside.

(in this analogy, two white spheres would be a normal, one white and one black sphere would represent a guaranteed het).

Now, the rules are, you only have one way to figure out what is in the bag. You reach in, take out one marble, look at it, put it back in the bag, and give the bag a shake.

You say when i pull out a white sphere 10 times, the chance for there being a black sphere in the bag is still at 50% and will forever stay at 50%. I say the chance for there being a black sphere inside is now at 0,1%.

it depends: do you disregard the information gained from repeatedly getting a look at one of the spheres? This is necessary to keep the value at 50% forever, no matter how often you try, all that data must be ignored. If its taken into account, and all available information is used, none ignored, then the bag, after 10 unsuccessful tries will no longer have a 50% chance of having a black marble inside.


The true chance of there being a black marble inside, after 10 misses in a row, is determined by answering the question: If there is a black sphere inside, what are the chances to miss out 10 times in a row and only see white spheres? AND THAT IS NOT 50%, it just isnt. Yes the bag with the marbles was initially purchased as a "50% possible het for black marble", but that doesnt matter, 50% is not the correct answer to the question, the correct answer is: slightly below 0,1%. or more precise: 0,5^10.
Put one white marble to represent normal,and one black marble the represent het in the bag. Now start pulling!