Jonathan Leland begins his 2010 paper “The hunt for a descriptive theory of choice under risk—A view from the road not taken” for the Journal of Socio-Economics.with a confession: his life’s path could have turned on trivial choices. A missed concert or an elective taken on a whim, and suddenly he’s not writing about decision theory at all, but about why money exists. His point is sly but devastating: tiny decisions cascade into whole lives. The same holds for economics itself — the field’s choice of how to represent risky gambles shaped an entire century of theory.

Economists, following von Neumann and Morgenstern, chose to describe lotteries as prospects: lists of outcomes with probabilities attached. That single decision — arbitrary in its own way — locked the discipline into obsessing over the independence axiom. Allais’ paradox, Prospect Theory, Machina’s triangles — decades of ink spilled trying to explain why people prefer certainty one moment and gamble recklessly the next. This was the road taken: technical, elegant, but narrow.

The road less traveled was Savage’s. He reframed lotteries as matrices of actions and states. Seen this way, different cracks appear: people violate transitivity and dominance, even preferring stochastically inferior options. From here, Loomes and Sugden developed Regret Theory: people don’t just weigh outcomes, they anticipate the sting of what they could have had. It’s FOMO formalized. In this light, seemingly “irrational” behavior — ripping a sealed booster box of trading cards instead of keeping it pristine — is the rational attempt to dodge regret.

But Leland presses further, down what he calls the road not taken. Here, the problem isn’t that people have quirky preferences. It’s that choices don’t reliably reveal preferences at all. Instead, choices reveal the rules we use to muddle through: crude similarity judgments. Does $10 feel much different from $9? At high probabilities, no — so people pocket the sure thing (“nothing to gain by gambling”). At low probabilities, yes — and suddenly it looks like “nothing to lose.” From this simple mechanism flows the famous four-fold pattern of risk preferences that Kahneman and Tversky codified.

This is where Leland breaks the spell. The holy grail of “preference-revealing choices” has been a mirage. Our so-called anomalies — Allais’ paradox, framing effects, addiction spirals — are not bugs but natural consequences of how we compare options. Addiction in particular becomes less a rational “self-medication” and more a vicious framing loop: each instance of consumption feels trivial, a choice where there’s “nothing to lose,” until the trap has already snapped shut.

So what to make of all this? Expected Utility is clean but ultimately naïve, Prospect Theory is dazzling but patchwork, Regret Theory gives us insight into FOMO, but Leland’s similarity judgments finally admit the obvious: we are creatures of context, not axioms. Our decisions do not unveil timeless preferences. They reveal the slippery heuristics we deploy when staring down uncertainty, haunted by happenstance and tempted by framing.

In the end, the real road not taken may be this: abandoning the fantasy of rational agents altogether, and building theories that take human messiness seriously. Leland, at least, had the nerve to step off the main path — and once you’ve seen what he’s pointing at, it’s hard to unsee.