I ran an Excel file on: "If an IQ test measured Intelligence reliably, how would it translate to a 3d6?" Here is what I got: (A plus or minus by a number means that the range generally extends beyond 0.4 past the number in that direction. Note that a score 100 is the boundary between 10 and 11 by definition.)

3d6	Min IQ	Max IQ
3	-45	61
4	-62	68+
5	69	74+
6	75	80
7	-81	85
8	-86	90
9	-91	95
10	-96	99+
11	-101	104+
12	105	109+
13	110	114+
14	115	119+
15	120	125
16	-126	131
17	-132	138+
18	139	155+

Given the classes of mental deficiency based on IQ (though malleable due to other factors), an intelligence score of 2 could correspond with an IQ between 20 and 44 inclusive and a score of 1 would be IQ<20. Conversely, IQ=180 could be the boundary between an intelligence "score" of 19 and 20 (with the boundary going down).

Still, as said in the previous comment, 5 Standard Deviations is likely the most precise level we can reach given a testable population of 1-2 billion. Therefore, we could define people with ImprobablyLowIQs / ImprobablyHighIQs as persons whose intelligence scores are too low/high to roll without modifiers.

BTW, my Intelligence score, as "rolled" in the 4th Grade, is a 15. Also, one has to wonder if a significant faction of MENSA are D&D players, as the boundary for membership corresponds to the lower limit of a 3d6=17 on the above table.

Most critics of IQ are those with a political agenda or view, without any technical credentials. The oft cited Mismeasurement of Man was written by Stephen Jay Gould, a paleontologist. The Making of Intelligence is a better critique as it is written by Ken Richardson, a developmental psychologist. Still, he is not a psychometrician.

This may be a factor of the True Believer Syndrome, but that is bound to happen in any field.

Among professionals in the field, some things are "beyond significant technical dispute", as described in The Bell Curve:

1. There is such a thing as a general factor of cognitive ability on which human being differ.

2. All standardized tests of academic aptitude or achievement measure this general factor to some degree, but IQ tests expressly desiged for that purpose measure it most accurately.

3. IQ scores match, to a first degree, whatever it is that people mean when they use the word intelligent or smart in ordinary language.

4. IQ scores are stable, although not perfectly so, over much of a person's life.

5. Properly administered IQ tests are not demonstratably biased against social, economic, ethnic or racial groups.

6. Cognitive ability is substantially inheritible, apparently no less than 40 percent and no more than 80 percent.

That doesn't mean that there is no evidence or argument to the contrary, but the concensus is reflected in the above statements.

Training will increase your IQ, but stopping the training will generally cause your IQ to backslide. While not a perfect analogy, it can be compared to physical fitness.

And training has its limits. I used to teach LSAT, as well as SAT on occassion. The questions do not test knowledge or technical proficiency in language or mathematics but rather thinking processes. Preparation courses can only boost your results by about half a standard deviation unless you are dealing with a particularly smart person who just doesn't "get it" but has an epiphany during the course.

It is important to note what IQ doesn't measure.

It does not measure one's ability to detect reliably or emote convincingly social cues. In short, one's charisma is on a different scale than one's intelligence. It is almost a trope in itself that many super smart characters will be socially awkward, perhaps suffering from Asbergers, such as The Big Bang Theory and Sheldon Cooper.

IQ is not a measure of drive, willpower, perseverence, moral fortitude and other such characteristics.

It is not a measure of "common sense" or wisdom.

One example of these contrasting traits is in military leaders. Patton, Napoleon et al. were smart men, but well below that of genius level, despite being known as "tactical geniuses".

Politicians are a less extreme example as several would qualify as geniuses, but their charisma and instincts (and money) is what allowed them to reach the highest offices.

"Social Intelligence" or "Emotional Intelligence" have been postulated to fill in these gaps. I have read allegations that they are better predictors of life outcomes than IQ (which has a statistically significant correlation, but not strong enough to predict the fate of any individual) but I have not seen anything remotely similar to the Bell Curve regarding these claims.

Additionally, IQ is not a measure of the abiity to understand complex (and even chaotic), multivariate systems. Richardson uses the example of professional horse race gamblers who are able to consistently pick winning ponies while only having normal I Qs. At the other end, one only has to look at how the government technocracy continually balls up the economy, war, social policy and such by putting the smartest guys they can find in charge.

Everything Bad Is Good For You is a book that discusses the increasing sophistication of popular culture and how it is making us 'smarter'. The author postulates that this could be at the core of the Flynn Effect.

Certainly, your average person is exposed to far more complex systems at an earlier age than people born more than a century ago. Take a 4 year old kid and one of their grandparents and give them a video game to play: who do you think will win?

But note that the Flynn Effect only happens within the average range. Goethe could still run mental circles around today's geniuses. But IQ is all about averages and deviations therefrom.

What is important about IQ scores is not the scores themselves, but the Z-scores that tell you how many standard deviations from the mean. One standard deviation makes you smart, two makes you MENSA smart, but it takes 3 standard deviations from the mean (or 145 for an SD of 15) to be a genius. Five SD is less than one in a million and 6 SD is about one in a billion, although it is hard to find calculators that can give you these numbers. Suffice to say, anyone with an IQ of near 200 has good cause to claim to be the smartest person on the planet (of 6 billion people).