A third of an SD of buffer over the line
Someone who tests 130 once can easily land 127 next time. At 135 there's room for normal day-to-day variation. Pick an observed score to see how often a retest would fall below 130.
Only supervised, normed, full-length tests count: WAIS, Stanford-Binet, Cattell or Mensa's own. A 135 from a casual web quiz isn't the same thing.
How many near-peers (130–140) are in the room?
The entry to the band elite achievers come from
Jonathan Wai's analyses of US elites (top executives, federal judges, billionaires, members of Congress) find them drawn heavily from the top 1% of ability, densest toward the very top. SMPY, followed for decades by Lubinski and Benbow, found top-1% adolescents later earning patents, doctorates and tenure far above base rates, and differences within the top 1% still predicted outcomes.
Not because other paths close, but because selection is highest here
The tail is where norming is thinnest
Tests are standardised on samples with relatively few people above +2 SD, so the raw-to-IQ conversion is less precise at 135 than near 100. Two good tests can disagree by several points up here. See how many people a norm sample actually has at this level.
Each term is the previous term multiplied by its own position number, counting from 1. The first term is 2. The sequence so far is 2, 4, 12, … What's the fourth term?
Percentiles, rarities, retest odds and norm-sample counts are computed from normal distributions. Elite-occupation, SMPY and Flynn-effect points summarise published research, not this site's data.