
Normal distribution with mean at 100 and standard deviation of 15.
The range is theoretically infinite, with the actual min and max IQ values determined by the current global population. But of course you knew that, being mathematically literate and all.
( , Wed 10 Dec 2014, 0:41, Reply)

and since points can only be assigned in the tests, not subtracted, its range has
You can describe data as being normally distributed even if it's impossible for points to fall above or below a certain limit, like finish times in a race or something.
( , Wed 10 Dec 2014, 1:21, Reply)

If race times are normally distributed around 28 minutes with a standard deviation of something or other, or bowling scores were normally distributed around 180, you wouldn't conclude that if you had enough people someone could theoretically get a negative race time, or a bowling score of -10 or 310.
edit: I see now- you're right :) The formula I was thinking of is the very old one, which is like (individual score/avg. score* 100). When SD is part of the formula rather than just an assumption about the data you can totally get a negative score.
( , Wed 10 Dec 2014, 13:07, Reply)

You need to be 6.666666... standard deviations below the mean, which er. means that you would need around 10^11 people for there to be 1 person with an IQ of zero. That's what wolfram alpha tells me anyway.
There's supposed to be about 7 billion now, so you might have to wait a few years for humanity to breed up to the point where you can achieve your dream. We might need a few more planets.
( , Wed 10 Dec 2014, 0:51, Reply)

So, a test to determine the IQ of an average person is not going to be precise enough to distinguish between the IQs of geniuses.
Presumably you can specifically test for someone with an IQ of zero, by asking them to take a test, and them not doing anything at all because they were born with anencephaly.
( , Wed 10 Dec 2014, 10:06, Reply)