So, as you know, if you flip a coin, you may expect a 50/50 distribution, there are two outcomes with probability equal to .5. You may take the exact odds of any contest, a roulette roll for an outcome, or blackjack, it doesn't matter. My question is this: How many rolls (or "events") does it take to expect the distribution to take shape to a measurable extent with a certain probability? For instance, could you expect the distribution to be 98% reflective after 200 events, with 95 percent confidence? Let me know if you have any questions.
-
-
I could be totally off track with my rudimentary statistics knowledge (just finished my first statistics class in college,) however, I tired to solve the problem with a 1-Prop Z-test on my TI-83. Again, this is probably completely wrong I just hope it can provide a start to solving the problem.
Here is what I did:
Null Hypothesis: distribution is 98% reflective after 200 events
Alternative Hypothesis: distribution is not 98% reflective after 200 events
95% significance level
---------------------
In my TI-83 I entered the following into 1-PropZTest
p: .5
x: 196
n: 200
prop: not equal to Po
I graphed that and got:
z=13.5765
p=0
So, since P< sigma we would reject the null hypothesis that the distribution is 98% reflective after 200 events.
-----------------
My answer just feels wrong as if I'm not correctly answering your question but hopefully it will provide a start. Please don't hate too much if it is wrong. Like I said, I only have a basic understanding of statistics and hypothesis testing WAS a sticky point for me throughout the second half of this past semester.Comment
