Tool
What to Watch Picker
Paste your watchlist, draw a shortlist of three, then commit to one.
What to Watch Picker
0 titles loaded.
Paste the watchlist you already keep somewhere, draw three at random, then commit to one of the three. Two stages, deliberately.
The single-answer version of this tool exists everywhere and nobody uses it twice, because a lone suggestion invites a re-roll and a re-roll invites another. Seeing three at once changes the psychology entirely: you are no longer accepting or rejecting a verdict, you are choosing between options, which is a thing people are willing to do.
The second stage is optional. Plenty of people draw the shortlist and then pick by feel, and that is a legitimate use — the randomness has already done the hard part by cutting forty titles to three.
How the What to Watch Picker works — and why it's fair
Stage one draws three distinct titles from your list using a partial Fisher-Yates pass: it swaps a randomly chosen entry into the first position, then draws the second from what remains, then the third. Because it draws without replacement, the same title cannot appear twice in a shortlist — a shortlist of three that contained two copies of the same film would be a shortlist of two.
Stage two draws one of those three uniformly, at an exact one-in-three chance each. Composing the two stages does not distort the odds: every title on your list ends up with an equal chance of being the final pick, because the shortlist itself was drawn uniformly. A forty-title list gives each title a 3/40 chance of making the shortlist and a 1/3 chance of winning from there, which multiplies to exactly 1/40.
That is worth stating precisely, because two-stage draws are a classic place for accidental bias to hide. If the second stage favoured the first-drawn entry, or the shortlist were built by taking the first three of a partial shuffle without re-randomising, the composition would no longer be uniform. Here it is, and the arithmetic above is the proof.
Nothing about your list is transmitted or stored. The titles you paste stay in the page; closing the tab discards them. That matters more than it sounds for a watchlist, which is a surprisingly personal document.
The tool deliberately does not know anything about the titles. It cannot filter by runtime, genre, or streaming service, because it has no data about your entries beyond the text you typed — and inventing that data by guessing at titles is exactly the kind of thing that produces confidently wrong suggestions.
When the What to Watch Picker is fair — and when it is not
What it does guarantee
- Every title on your list has an identical chance of being the final pick — 1 in n, exactly, with the two stages composing cleanly.
- No title can appear twice in one shortlist, so the shortlist is always three genuinely different options.
- The order the three appear in carries no information; the final draw is uniform across all three.
What it does not
- It has no idea what any of your titles are. A four-hour art film and a twenty-minute comedy are indistinguishable to it.
- Duplicated lines are two entries and get double the odds, which is occasionally what people want and more often a paste error.
- Drawing a fresh shortlist repeatedly until you like one converts this back into a manual choice — the tool cannot stop you, but that is no longer a random pick.
Two worked examples
A 40-title backlog, one shortlist
- Each title has a 3-in-40 chance of appearing in the shortlist, which is 7.5%.
- Each of the three then has a 1-in-3 chance of the final pick.
- Net: every title has exactly a 1-in-40 chance, or 2.5%, of being what you watch.
A 6-title shortlist someone else already made
- With six titles the shortlist is half the list, so the first stage barely narrows anything.
- Each title has a 3-in-6 = 50% chance of reaching the shortlist, then 1-in-3 from there.
- Net 1-in-6, 16.7% — at this size, drawing the final pick directly would be simpler.
Where the two-stage version earns its keep
The classic case is the couple who have spent twenty minutes scrolling and are now annoyed with each other rather than with the interface. A shortlist of three gives both people something to react to without either having proposed it, which removes the thing that was actually causing friction.
It also works well for a group where one person's taste tends to dominate. Because the shortlist comes from the combined list and nobody chose it, the argument shifts from whose suggestion wins to which of three neutral options is best — a much shorter argument.
Film clubs and watch-along groups use it to break a rotation that has quietly calcified, where the same three members keep proposing and the rest keep agreeing. Feeding everyone's suggestions into one list and drawing from it changes who gets represented without anybody having to say so.
If you want the ceremony rather than the arithmetic — a group watching a spin land — put the titles on the spinner wheel instead. If your list is long enough that some titles genuinely deserve better odds than others, the weighted random picker handles that properly.
Frequently asked questions
Does the shortlist stage change each title's overall odds?
No. A 3-of-n shortlist followed by a 1-of-3 pick multiplies out to exactly 1-in-n, so every title has the same chance as it would from a single direct draw.
Can the same title appear twice in one shortlist?
Not unless you pasted it twice. The shortlist is drawn without replacement, so three distinct list positions are always chosen.
Can I draw a shortlist of five instead of three?
Three is fixed. Beyond about four, the shortlist stops feeling like a decision and starts feeling like the original list again, which defeats the purpose.
Does it know anything about the films I list?
Nothing at all. It sees text, not titles — it cannot filter by runtime, genre, or where something is streaming, and it will not invent that information.
Is my watchlist sent anywhere?
No. It stays in the page and is discarded when you close the tab. Nothing is stored, uploaded, or logged.
What if my list has fewer than three titles?
The shortlist is capped at the list length, so a two-title list produces a two-title shortlist — at which point the this-or-that picker is the more honest tool.
Should I redraw the shortlist if I dislike all three?
You can, but be aware of what it does: repeated redrawing until something appeals turns this back into you choosing, with extra steps. One redraw is fine; four is a decision you are making by hand.
Does the order the shortlist appears in mean anything?
No. It reflects nothing but draw order and carries no ranking. The final pick is uniform across all three regardless of position.
Can I use this for something other than films?
Yes — it works on any list where you want a shortlist rather than a verdict: restaurants, books, weekend plans. The two-stage shape is the feature, not the subject.