It's 7:14 p.m., you've had the same delivery app open for eleven minutes, and you have scrolled past forty restaurants without choosing one. You're not undecided because the options are bad. You're undecided because there are forty of them, each with a rating, a wait time, and a faint promise that the next one down might be the right one. A random restaurant picker fixes this not by giving you more information, but by taking the decision out of your hands entirely โ and it works far better than it has any right to.
More options make the choice harder, not easier
We tend to assume that a longer list of restaurants is a gift. In practice it's a tax. Every additional option you consider raises the bar for what counts as "good enough," because somewhere in that list there might be a slightly cheaper, slightly closer, slightly better-rated place. Psychologists call the two modes maximizing and satisficing. A maximizer wants the best possible dinner and so must, in theory, evaluate everything; a satisficer wants a good dinner and stops at the first one that clears the bar. The maximizer eats later, eats more anxiously, and โ this is the cruel part โ usually ends up less happy with whatever they chose, because they spent the whole meal aware of the road not taken.
The infinite scroll is built for maximizers, or rather it manufactures them. It keeps the list moving so the bar keeps rising and the bottom never arrives. A random pick collapses all of that. It hands you one place, and one place is a decision you can actually act on. We dug deeper into this trap in why nothing sounds good to eat, but the short version is that the paralysis is a feature of the format, not a flaw in you.
Chance ends the regret loop
The real damage of a long deliberation isn't the time it costs up front โ it's the regret loop it sets up afterward. Choose a place from a ranked list and a small voice follows you to the table: but what if the one two spots down was better? You picked it, so you own the outcome, so any disappointment is your fault. That ownership is what makes self-directed choosing exhausting.
When chance picks the place, you stop auditing the decision and start enjoying the meal.
Randomness quietly removes that burden. Nobody weighed the options, so there's no better option you failed to spot. The place was simply dealt to you, like a card, and you play the hand. People who let a coin or an app decide consistently report enjoying the meal more โ not because the food is better, but because they've been freed from grading their own judgment all night.
Constrain the pool before you let chance in
Here's the catch that separates a good random pick from a bad one: randomness is only as good as the set it draws from. Spin a wheel that includes every restaurant in a thirty-mile radius and you'll eventually land on a gas-station taquito or a place that closed in 2019. The trick is to do the thinking once, at the level of the pool, and then never again at the level of the individual choice.
So constrain first. Set a radius you'd actually be willing to drive โ not the theoretical maximum, but the honest one for a Tuesday. Then either pick a cuisine you're in the mood for or leave it fully open if you genuinely don't care. And cut the chains, because a random pick that surfaces the same franchise you could've named from memory defeats the entire point; the magic is in being sent somewhere you wouldn't have thought of. Once the pool is good, every result inside it is good, and you can trust the dice. If you're stuck even on the broad direction, a few prompts for what to eat tonight can help you set the constraint before you randomize.
Re-roll once, then commit
A picker is only useful if you respect its verdict, and the failure mode is obvious: you keep tapping until it gives you the answer you secretly already wanted, at which point you haven't randomized anything โ you've just laundered your own indecision through an app. The fix is a rule, set before you start. Re-roll exactly once. The first pick gets a real moment of consideration. If it's genuinely wrong โ too far, wrong mood, you ate there Monday โ you may roll a second time. Whatever the second roll says, you go.
This works because the point was never to find the optimal restaurant. The point was to eat a good dinner without spending your evening deciding. One re-roll keeps a veto for the truly off pick while preventing the infinite tap-tap-tap that turns the picker back into the scroll you were trying to escape.
This is exactly what Tonight's Table does
If everything above sounds like the description of an app, that's because it is the one we built. Tonight's Table is a random restaurant picker with the constraints baked in. Set your radius up to forty-five miles, choose a cuisine or hit Surprise Me to go fully open, and flip on the hide-chains toggle so the result is always somewhere small and independent rather than a logo you already know. Then tap once. It surfaces a single nearby spot โ not a ranked list to second-guess, just one place to go.
Don't like it? Tap again, and remember the one re-roll rule. Mark places visited as you go so it stops sending you back to the same few, and over a few weeks the chain pulls you out of the five-restaurant rut most of us fall into without noticing. It pulls its data from Apple Maps, asks for no account, and is free to download โ built for the night you'd rather let chance choose than scroll for one more minute.