Strong band · 120+
You held accuracy where the rule stopped announcing itself. Simple series separate almost nobody — 2, 4, 6, 8 is a warm-up. The band opens on items where the rule changes mid-sequence, where a conclusion has to survive a negated premise (if P then Q, and Q is false), and where two sets overlap and the answer depends on not counting the overlap twice. Scoring here means those held. What it does not measure is reasoning under real conditions, where premises arrive incomplete and nobody marks which ones matter. It is a clean read on a narrow skill: following a structure you were never told, without losing the thread.
Mid-band · 90–119
You held the items where the rule was visible and lost ground where it had to be reconstructed. This is the most common shape on the scale, and the split is usually specific rather than general: arithmetic and letter series stay solid, then a sequence changes its rule halfway through, or a syllogism turns on a negation, and the answer that feels right is the one the structure rules out. Of the four scales this is the most trainable — the moves are a finite, nameable set. Most people gain by slowing down exactly where an item stops looking familiar, not by thinking faster.
Below sample · under 90
This scale rewards one narrow skill — following a rule nobody stated — and on this sitting it did not carry. Worth separating from what it is not: not general intelligence, not how well you think about things you already know, and not a fixed quantity. A sitting done tired, rushed, or on a phone in motion measures conditions as much as reasoning. It is also the most learnable of the four scales. Series, odd-one-out and simple syllogisms are a small closed set of patterns; people who spend an afternoon on them usually move this band on the next attempt, which tells you what the number is worth.