One of four reference posts, each 15 points apart
The scale is built on fixed posts: 70, 85, 100, 115. About 68% of people fall between 85 and 115, so 85 is the exact boundary of the central two-thirds. Step one point either side and the label changes, but almost nothing about the person does.
A full standard deviation separates them
The same distance as from 85 up to the dead centre. Modern scales separate low average (80–89) from borderline (70–79) for exactly this reason.
Where did your 85 come from?
The curriculum is pitched at the middle. Here's where it gets steep.
Pace, not capability, is usually the limit. Click a step.
Raw scores rose roughly 3 points a decade through the 20th century, with gains concentrated in the abstract reasoning schools reward; that's one reason tests are re-normed every 10–20 years.
Inside AFQT Category IV, with a cap on how many get in
Category IV covers the 10th to 30th AFQT percentiles, about IQ 81 to 92 on a normal curve. Federal law caps Category IV at 20% of recruits, and the services usually take far fewer. The caution traces back to Project 100,000 in the 1960s, whose low-scoring inductees had markedly worse training outcomes.
The bar reflects the cognitive load of technical military training, not an inability to serve. Many people at this level qualify and serve well.
Where people at 85 routinely do well
General ability predicts job performance better than most single factors, but it explains only part of the variance, and less as jobs get more concrete. In these roles, conscientiousness (dependability, persistence, follow-through) most separates strong from weak performers, and it's largely unrelated to IQ.
Percentiles, ranges and AFQT conversions are computed from a normal distribution (mean 100, SD 15). Military, job-performance and Flynn-effect points summarise published research and law; they aren't this site's data, and the 20% cap is the statutory ceiling, not current practice.