There is a version of you, on a Tuesday night, who has spent eleven minutes deciding where to eat and has not yet decided. You have scrolled three apps, opened two menus, dismissed a place because the parking looked bad, and reopened the first app to find the place you almost picked again. The food, wherever you end up, will be fine — it was always going to be fine. The eleven minutes are the only thing you actually lost. I want to make an argument that sounds reckless and turns out to be true: in a moment like that, you would be better off flipping a coin.
The math of "good enough" is not the math you think
Here is the part people skip. When you are choosing between dinner options that are all, honestly, somewhere between decent and great, you are not choosing between a triumph and a disaster. You are choosing between a 7 and an 8 on a scale you invented and cannot measure. The gap between those outcomes is small. The effort required to reliably find the 8 instead of the 7 is not small — it is the scrolling, the second-guessing, the menu-reading, the asking your partner who also doesn't know. The return on that effort, in actual enjoyment of your actual dinner, is a rounding error.
Optimizing only pays when the options are genuinely far apart and the stakes are real. Dinner on an ordinary night is the opposite: tightly clustered options, low stakes, and a steep cost to keep searching. Under those conditions the rational move is not to find the best one. It is to grab a good one and stop.
Maximizers are miserable; satisficers eat
Psychologists have a clean name for the two ways people approach this. A maximizer needs to be sure they got the best possible option, which means they cannot stop until they have considered everything — and even then they wonder about the thing they didn't try. A satisficer sets a bar for "good enough," takes the first option that clears it, and moves on with their evening. The research on this is unkind to the maximizers: across studies, they end up with objectively comparable outcomes and feel measurably worse about them. More options, more comparison, more certainty-seeking — and less satisfaction at the end.
This is the heart of what Barry Schwartz called the paradox of choice. Beyond a handful of options, each additional one doesn't expand your happiness; it expands the shadow of everything you didn't pick. Twenty nearby restaurants don't make dinner twenty times better. They make whatever you choose feel like it came at the expense of nineteen alternatives you'll quietly mourn.
The cost of a choice isn't the option you take — it's the nineteen you spend the meal wondering about.
Randomness deletes the regret you were about to manufacture
A random pick does something sneaky and wonderful: it removes the author of the decision. If you agonized your way to a restaurant and it's merely fine, that's on you — you should have known, you should have checked, you should have picked the other one. But if a coin or an app handed you the place, there's no counterfactual to torture yourself with. You didn't pass over nineteen better options; you were simply sent somewhere, and you went. The decision cost drops to zero and the regret has nowhere to attach. You are free to just enjoy a perfectly good plate of food, which was the entire point of going out.
There's a second gift hiding in the randomness, and it might be the larger one. Left to your own devices, you order the same four things from the same three places, because choosing is exhausting and the known quantity is safe. A random pick reaches past your defaults. It sends you to the Burmese place you'd have scrolled right past, the diner you assumed was nothing special, the spot two neighborhoods over you never had a reason to try. Some of those will become favorites — and you would never, ever have chosen them on purpose. Ruts don't break themselves. If you've ever stared at every option and felt nothing, the piece on what to do when nothing sounds good to eat is the companion to this one.
The honest catch: garbage in, garbage out
I have to be straight about the limit, because randomness is not magic and the internet is full of people pretending it is. A coin flip is only as good as the options it chooses between. If your pool is a strip of fast-food chains and a gas-station taquito, randomness will faithfully deliver you a bad dinner, no regret-deletion required. Random selection doesn't create quality. It only spares you the labor of picking among options that are already worth picking among.
Which means the real skill was never the final choice — it was building the pool. The work worth doing is upstream: deciding that you want nearby, independent-leaning places rather than the same four logos; setting a radius that's wide enough to be interesting and tight enough to be practical; quietly cutting the options you already know you don't want. Do that curation once, and the final pick genuinely doesn't matter. You've loaded the dice so that every face of them is good. If you want a structured way to think about that upstream step, how to decide what to eat tonight walks through it.
Why this is exactly what the app does
This is the entire design idea behind Tonight's Table, and it's why I keep coming back to the coin-flip framing. The app doesn't roll the dice on the whole world — it rolls them on a pool that's already been curated for you. Everything it considers is nearby. It leans toward small, independent places, and if you'd rather not see chains at all, one toggle hides them. You pick a cuisine or hit Surprise Me, set a radius up to forty-five miles, and tap once. It hands you a single place. Not a ranked list to re-optimize against — one place, with the regret-machinery already disconnected.
If the pick is too far or genuinely not the mood, you tap again and it re-rolls; mark spots visited so it stops sending you where you've already been. That's the whole loop, and it's faithful to the argument: curate the pool well, then let randomness spend your decision-budget on absolutely nothing. Tonight's Table is free to download, needs no account, and exists to take the eleven minutes back.