A recent paper, "Binding-Motivated Contextuality," proposes running quantum-foundations-style cyclic tests — the kind used to prove that some physical systems can't be explained by any single, order-independent classical model — on ordinary human probability judgments. Its own version of the test is preregistered but hasn't been run yet. This page runs a simplified stand-in on you, right now, using same-game-parlay props instead of lab stimuli.
You'll give 8 quick gut-probability judgments about a hypothetical basketball game. The catch: the same handful of correlated propositions get revisited in different pairings, order, and framing. Real classical probability theory — just Bayes' rule, nothing exotic — puts a hard, unforgiving constraint on how those 8 answers must relate to each other if you're internally coherent.
Almost nobody is. Let's see your number.
Four things about this game are tangled up with each other. You won't be asked about all four at once — just pairs, one at a time, in an order that's reshuffled every time this page loads.
Answer fast, on instinct — like you're setting a live in-game number, not doing homework.
Bayes' rule says, for any two events X and Y: P(Y|X)·P(X) = P(X and Y) = P(X|Y)·P(Y). Chain that identity all the way around a closed loop of four propositions (A→B→C→D→A) and every marginal probability cancels out. What's left is one bare, unforgiving requirement: multiply the four conditional probabilities going one way around the loop, multiply the four going the other way — if you hold one single coherent belief about this game, those two products must be equal. Not close. Equal. No quantum mechanics required, just a closed loop and Bayes' rule. This is the plain-probability cousin of the cyclic KCBS / Suppes–Zanotti inequalities the paper borrows from quantum contextuality tests, and of the "nonzero class in cohomology with no global section" language the paper uses for why binding and contextuality share a mathematical signature.
—
Same-game parlays are priced leg by leg, in an order the book chooses, and each leg's price is really a conditional probability on whatever's already "in the slip." If your gut probabilities for a cyclic set of correlated props are order-dependent — which the test above almost certainly just showed — then the "fair" combined price of a parlay depends on which order you were shown, or happened to think about, the legs. That's not a metaphor for irrationality. It's the same broken loop, in dollars. The book doesn't need you to misjudge basketball. It just needs your judgments to be non-commutative. They already are.