What should you watch tonight?
What should we watch?
Every slice is the same size
1 in 10
Press SPIN to draw a winner
Seven dinners, drawn
WeeklySeven dinners for the week ahead, spun fresh every Friday. Deciding what to eat is the tiring part, not cooking it.
Try also
How do you decide what to watch without scrolling for an hour?
Spin for the genre, not the title. The reason picking a film takes forty minutes is that streaming menus ask you to choose from thousands of things at once, and nobody can hold thousands of options in their head. Narrow it to one genre first and the same menu becomes a short list you can actually read.
Ten genres on the wheel is a 10% chance each. The winning slice comes from your browser’s cryptographic random number generator, so nobody chose it โ which is exactly what makes it acceptable to everyone in the room. Then open whichever service you have, filter to that genre, and take something from the first screen.
Genres are the right unit because they do not go out of date. A wheel of specific titles is stale within a season and wrong the moment something leaves a service. Comedy, thriller and documentary will still be there in five years, and the wheel will still work when the streaming landscape has rearranged itself twice over.
How many slices to use depends on how much of the evening you are prepared to spend deciding.
| Slices | Chance each | Best for |
|---|---|---|
| 2 | 50% | The last two contenders, settled |
| 4 | 25% | One per person, everyone represented |
| 6 | 16.7% | The genres this household actually watches |
| 10 | 10% | The default: a real spread, including things you would not have picked |
| 15+ | under 7% | A month of film nights with remove-winner on |
The four-slice version is worth a mention on its own. Have everybody add one genre, then spin: every person is represented, nobody has to argue out loud for the thing they want, and the result has a mandate from nobody at all โ which turns out to be the most acceptable mandate available.
For the other half of the evening, the what to eat wheel settles dinner, and if it comes down to a straight two-way call the yes or no wheel is faster.
Frequently asked questions
Why spin for a genre instead of a film?
Because a genre is a decision you can make in one second and a film is not. Choice overload is well documented: past a certain number of options people take longer, enjoy the result less and often defer entirely โ which is the hour of scrolling. Cutting thousands of titles to one genre restores a list small enough to actually choose from.
Can I put streaming services on the wheel instead?
Yes, and for households with three or four subscriptions it often works better. Type one service per line, spin, and browse only that one. You can also mix the two โ put your services on the wheel, spin, then spin again on a genre wheel. Two seconds and the search space has shrunk by two orders of magnitude.
Can I add my own genres or moods?
Type them on their own lines. Every line becomes one equal slice, so the odds are just your line count. Mood works as well as genre โ ‘something funny’, ‘something we have seen before’, ‘something with subtitles’ โ and for a lot of people the mood is the real question the genre was standing in for.
How do I make some genres more likely?
Add the line twice. Because each line is one equal slice, a genre listed twice on a twelve-slice wheel has a 2 in 12 chance, and everyone watching can count the slices. That visibility is the point โ a hidden weighting slider would make the wheel useless for settling anything with another person in the room.
What does remove-winner do?
It takes the chosen genre off the wheel and redraws, so the next spin picks from what is left. It is how to run a whole week: spin on Monday, watch, and the wheel will not hand you the same genre again until you reset it. For a single evening leave it on and it also stops the wheel repeating itself if you spin twice.
Is anything I type sent anywhere?
No. Your list lives in the browser and in the page address, and the spin runs locally. Close the tab and it is gone. Copying the link shares the exact wheel you built, which is also the only way to keep one โ bookmark the page after editing and it reloads with your genres in place.
How the randomness works
crypto.getRandomValues using rejection sampling, so a ten-item list gives every item exactly a
10% chance with no bias towards the first entries. Duplicate lines are treated as separate slices,
which is how weighting works. With remove-winner enabled the chosen genre is removed from the
working list and the wheel redraws at equal widths, raising the odds for everything left. Your
list stays in the browser and in the page URL; nothing is sent anywhere and no history is kept
between spins. The default list is genres rather than titles deliberately: genres do not expire
when a licensing deal does.Sources:
Why the scroll is the problem, not the catalogue
Streaming did something odd to watching television. The catalogue got bigger every year and the satisfaction of choosing from it got smaller, because the two things are related: past roughly a dozen options, adding more makes the decision harder rather than better. Psychologists call it choice overload, and the streaming menu is close to a laboratory demonstration of it.
A wheel does not solve that by being clever. It solves it by removing the part that hurts โ the responsibility. Once the genre is chosen by something with no opinion, nobody in the room is advocating for anything, nobody has to be talked round, and the fifteen minutes normally spent reading the same six thumbnails go back into the evening.
The one rule that keeps it working is the same as with any decision wheel: whatever comes up, you watch. A wheel with a veto attached is a suggestion engine, and a suggestion engine is what the streaming menu already was.