It is half past nine on a Tuesday. You have ninety minutes before bed, a cup of tea going cold, and a decision to make about how to spend the evening. The question that should interest anyone who studies online casino behaviour — but also anyone who studies decision-making at all — is why the size of the menu changes what we actually do, and not always in the direction we assume. Conventional wisdom says fewer options mean faster decisions; the evidence on follow-through suggests something stranger, particularly when the choice is made late and under mild fatigue.

The paradox: more options, better completion

The standard story about choice overload comes from Iyengar and Lepper's 2000 jam study: shoppers presented with 24 varieties were less likely to buy than those shown six. That finding lodged itself into product design, onboarding flows, and the received wisdom of anyone who has ever built a menu. But the jam study measured purchase, not persistence. Once a choice is made, the picture changes.

A less-cited pattern shows up when you measure what happens after selection rather than at the point of selection. Smaller initial menus produce quicker decisions but weaker commitment; broader menus slow the choosing down while producing people who are more invested in what they picked. In a 2019 field experiment on a subscription service, users offered twelve onboarding paths completed their first week at a measurably higher rate than users offered four — despite spending longer on the selection screen and reporting more difficulty choosing.

The mechanism is not mysterious once you look at it through the lens of commitment. Choosing from four feels like picking a default. Choosing from twelve forces a comparison, and comparison generates reasons. Reasons, in turn, are what we use to justify following through when the initial enthusiasm fades an hour later.

Why 9pm changes the arithmetic

Late evening is a specific cognitive environment, and it matters more than most people acknowledge. By nine o'clock, most adults have been making decisions for fourteen hours. The prefrontal machinery that supports deliberate comparison is running on fumes, while the affective system — the part that responds to novelty and immediate reward — is still fully awake.

This is where the twelve-versus-four distinction gets interesting. Under low cognitive load, a four-option menu is trivially easy and produces near-instant selection. Under the fatigue of a long day, a twelve-option menu does something different: it recruits what remains of your deliberative capacity into a brief but genuine evaluation. That evaluation, short as it is, creates a small sense of authorship. You didn't just tap the obvious thing; you weighed alternatives and arrived somewhere.

Kahneman's framing of System 1 and System 2 is useful here, though it is often over-applied. The point isn't that late-night choices are irrational. It's that the type of processing available at 9pm is different from the type available at 11am, and menu size interacts with that difference. A short menu invites System 1. A longer menu, even a tired System 2, produces a trace of reasoning that later supports follow-through.

The fatigue trap in reverse

There is a counterintuitive corollary. The people most likely to abandon a chosen activity are not those who chose quickly — they are those who chose too easily and then met friction. A twelve-option menu, by slowing the decision by perhaps fifteen or twenty seconds, front-loads a tiny amount of friction at exactly the moment when friction is cheap. Four options defer all the friction to the moment of execution, when it is expensive.

Reward loops and the shape of the first hour

Behavioural psychology has a well-established account of what keeps people engaged once they start: variable-ratio reinforcement, the intermittent schedule that produces the most persistent behaviour. Skinner's work on this is foundational, and its reach extends far beyond any single domain. But variable-ratio schedules explain maintenance, not initiation. They tell you why someone keeps going once the loop is running.

The question of initiation is different, and it is where menu size does its work. A decision made from a broader set produces a stronger initial commitment, and that commitment carries the first hour. After that, the reinforcement schedule takes over. The two mechanisms are sequential, not competing: breadth at the point of choice, variability at the point of repetition.

There is a practical implication hiding in this sequencing. If you want someone to still be engaged at 10:30pm, the design question at 9pm is not "how do I make this easier?" It is "how do I make the initial choice feel like it was theirs?" Twelve options do that better than four, provided the twelve are genuinely distinct and not twelve shades of the same thing.

Loss aversion and the cost of a wrong pick

Kahneman and Tversky's loss aversion — the finding that losses loom roughly twice as large as equivalent gains — is usually invoked to explain why people hold losing positions too long. It applies just as well to how people treat a chosen evening activity.

When you pick from four, the cost of being wrong is low and the counterfactual is obvious: the other three were right there. When you pick from twelve, the counterfactual is diffuse. You cannot easily imagine the eleven alternatives you didn't take, so the chosen option acquires a kind of default status. Abandoning it feels less like switching and more like waste.

This is not unambiguously good. A diffuse counterfactual can trap people in a choice that isn't working. But at the ninety-minute scale of an evening, the effect is mostly benign: it provides just enough stickiness to get past the first ten minutes, which is where most abandonment happens anyway.

A concrete illustration

Consider two groups given the same task — learn a short piece of software, complete a small creative exercise, whatever the evening's activity happens to be. Group A picks from four labelled paths. Group B picks from twelve. In the first five minutes, Group A is faster and reports higher satisfaction with the choosing process. By minute thirty, Group B has caught up on satisfaction and is ahead on completion. By minute sixty, the gap is wider still. The pattern holds across several replications with modest effect sizes — not dramatic, but consistent, and in the opposite direction from what the jam study would predict.

What this means for how we design the evening

The forward-looking question is not whether twelve beats four in general. It is under what conditions breadth helps, and late evening is one of them. Three conditions seem to matter: the options must be genuinely distinguishable, the choice must be made once rather than repeatedly, and the follow-through window must be short enough that the initial commitment is still warm when execution begins.

Where those conditions hold, a broader menu is the better instrument. Where they don't — where options blur together, where the choice is revisited every few minutes, where follow-through is hours away — the jam study's logic reasserts itself and brevity wins.

The interesting research direction is not more menu-size studies but the interaction between time of day, menu breadth, and commitment strength. We know a great deal about how people choose and rather less about how the choosing shapes what happens next. The 9pm window, with its particular mix of fatigue and resolve, is as good a place as any to start looking.