PDA

View Full Version : Finding the "bell" using Binomial Distribution

AceKQJT
11-01-2004, 12:48 PM
Ok. So I'm gonna flip a coin 1000 times. I am looking for a range in which heads comes up with 90% certainty. I'll give an example, because I can't even understand what I've written:

I know that the chance of heads coming up exactly 500 times is 5.253%. I can calculate the binomial distribution for each case of "successes", then manually find that 90.6% of the time, heads will show up between 474 and 526. I could express this by saying

"If you flip a coin 1000 times, there is a 90.6% chance that the coin will come up heads 500 (+/- 26) times."

Is there any way to come to this result other than calculating individual successes for the entire trial, and then manually adding the results?

--Casey

BlueBear
11-01-2004, 12:55 PM
Yes, my statistics is hopelessly rusty but I think for a large enough sample size (such as 1000), you can use approximate the binomial distribution as a normal distribution and calculate these results easily. Any basic statistics text will discuss this in greater depth.

Deli
11-01-2004, 05:31 PM
Have a look at this site (http://faculty.vassar.edu/lowry/binomialX.html) that lorinda linked to recently.

AceKQJT
11-02-2004, 10:24 AM
Yeah...I can come up with the numbers this site does. I guess I'll just figure that +/- 2xSD will give me roughly 97% certainty.

Thanks anyway,

--Casey