Question Wheel

Running this wheel for a class, a meeting, or a household? The guide Using the wheel for groups: classrooms, meetings, households covers the group mechanics — loading a roster, removing duplicates, eliminator mode so nobody is picked twice, sharing a prepared wheel with a co-host, and resetting afterwards. This page is about the question preset itself: what the 25 questions are, how the circle is divided between them, how one spin picks a question, how the parser reads a question you type, and how to run the deck without repeats. Every guide is listed at /guides/.

Why the wheel asks better than you do

The same person picking the icebreaker every week produces the same flavor of question every week — and the group learns to pre-compose answers, which defeats the purpose. A random draw keeps questions genuinely surprising, distributes the awkward ones fairly, and removes the facilitator's thumb from the scale. The moment of communal suspense while the wheel slows down is not nothing, either: it gets a room's attention the way "so, icebreaker time" never has.

Everything below is a walk through what the engine actually does with this preset. Every angle, percentage and count on this page is computed at build time by the same code that runs the wheel, and the worked spins use scripted bytes so they can be replayed — a live spin draws its bytes from crypto.getRandomValues instead. The full mechanism is on the methodology page.

What the preset is, and how to edit it

The deck is 25 original questions written for this site, loaded in list order into the same editable entry list every wheel on the site uses. They are ordinary entries: one question per line, at most 80 characters each. The longest shipped question runs 59 characters, so every one of them fits with room to spare. Because the wheel packs 25 labels into one circle, each label is trimmed on the wheel itself; the full question always appears in the result box, and that is the line to read out. The shipped questions are calibrated for mixed company — colleagues, students, relatives at the same table. For a closer-knit group, swap in spicier ones; for a classroom, seed it with revision prompts and let the wheel cold-call topics instead of students. Delete the questions your group has outgrown, paste in your own, or mix the two. The edited list is saved in this browser's local storage so it survives a reload, and Reset list restores the shipped 25. A wheel holds at most 100 entries, and the cap is announced rather than applied: paste 5 copies of the deck and the engine refuses with "A wheel holds at most 100 entries — this list has 125. Remove some entries; nothing was silently dropped."

25 segments of 14.4° each

A wheel of N entries is divided into N equal segments of 360/N degrees, in list order, with every boundary computed as an exact multiple. For this deck N is 25, so each question occupies a 14.4° segment and wins 4% of spins. The first question, "What tiny purchase improved your life the most?", starts at 0° clockwise from the pointer and ends at 14.4°; entry number 13, "What is the furthest you have traveled for something silly?", is centred at 180°, directly opposite the pointer. When a question wins, the wheel is rotated so that segment's midpoint stops under the pointer — the rotation is worked out from a winner that already exists, not the other way round. Colours cycle through a fixed 8-colour palette, and this deck happens to be a size that needs the cycle's one adjustment: with 25 segments the last would have come round to the same colour as the first, so the engine gives the last segment the second palette colour instead. The geometry exists for the eyes only. A 14.4° segment here is exactly as likely to win as a half-circle on a 2-entry wheel of equals, because the pick is made from random bytes before any geometry is involved; the segment geometry guide covers the arithmetic at every wheel size.

One scripted spin, byte by byte

Two rules are visible in that one spin. The mask: the engine takes the minimal number of bits whose range covers 25 and ignores the rest of the byte, which is why bytes 7 and 231 — different only in their top bits — both produce "What sound instantly relaxes you?". The rejection: a masked value at or beyond 25 is discarded and a fresh byte drawn; the engine never takes a remainder. To show the difference that makes, the build runs all 256 values a byte can hold through uniformInt as the first byte of a spin. The engine accepts 200 of those on the first draw and rejects 56; the accepted bytes land on the 25 questions exactly 8 times each. Taking the byte modulo 25 instead would have kept every byte and handed the leftovers to the questions at the top of the list — that is modulo bias, and the redraw exists to prevent it. Because the mask keeps acceptance above half, redraws are rare and the spin stays instant. The pick is made the moment you press Spin; only then does the roughly four-second ease-out rotation begin, calculated to decelerate onto the segment that already won. If your system asks for reduced motion, the rotation is skipped and the identical question is announced at once. The guide How the wheel picks a winner has the full walk-through.

Running the deck without repeats

In normal mode each spin is an independent draw over all 25 questions at 4% each, and nothing stops a question from coming up twice in one session. Tick eliminator mode and each spin removes the question it picks: the elimination order becomes the order the questions get asked, and no question is asked twice until you reset. Each elimination is a fresh uniform draw over the questions still on the wheel — 1/25 on the first spin, 1/24 on the second, and so on. The walk-through below runs a 5-question warm-up, the first 5 questions of the deck, with scripted bytes; every row is the engine's own output.

Spins 1 and 3 needed 2 and 2 draws each because a scripted byte fell outside the shrinking range and was rejected — the redraw rule applies at every step, not just on the full wheel. After spin 4, one question remains, "Which app on your phone would be hardest to give up?", and the engine marks the run finished rather than spinning for a single entry: ask it, then reset. Spinning a finished run answers "This elimination is finished — reset to run it again." Because each spin is uniform over what remains, every one of the 5 questions ends up in that last slot with probability exactly 1/5, or 20%.

The same mechanism scales to the whole deck. The build runs the full 25-question eliminator with a scripted tape: 24 spins ask 24 different questions — the first out is "Which invention are you shocked does not exist yet?" — and leave "What do you always overpack and never use?" as the final, unspun question. The byte source was consulted 27 times along the way, so 3 redraws happened and changed nothing about fairness. Two rules follow from the mechanism. Elimination states are immutable values, so a run in progress is replayable step by step and only a spin can advance it — but it exists only on the page and ends with the tab, while the list is what the browser saves. And editing the list mid-run restarts the elimination on the new list. For a standing meeting that must not repeat across weeks, delete the asked questions from the list itself. The spin history keeps only the most recent spins, so on a longer deck note the asked questions down as you go rather than relying on the history at the end. The guide Eliminator mode, explained covers the rest.

Your own questions: lines, commas and the length cap

  Coffee, tea, or neither?

What would you never lend out?
Mountains, beach, or city for a week off?   

The length cap is honest in the same way. A question of 120 characters exceeds the limit of 80, and the parser refuses the whole list with "Entries are limited to 80 characters — "If you could permane…" is too long to fit on a segment." rather than clipping it. Shorten it or split it into two questions. Duplicates are kept unless you say otherwise: the same question on two lines, one of them in capitals, plus a third question parses to 3 entries, and ticking Remove duplicate entries collapses it to 2, keeping the first. The full rule set is in Entry lists and parsing rules.

Sharing a deck by link

Copy share link percent-encodes each question and joins them with | into a #items= fragment on this page's address. The 3-question list above encodes to Coffee%2C%20tea%2C%20or%20neither%3F|What%20would%20you%20never%20lend%20out%3F|Mountains%2C%20beach%2C%20or%20city%20for%20a%20week%20off%3F — the commas travel as %2C and the spaces as %20 — and decoding it returns the 3 questions unchanged, commas and all. The shipped deck of 25 questions encodes to 1699 characters, against a cap of 1800, so the whole deck fits in one link; the full address runs 1747 characters. Browsers do not include the fragment in web requests: a co-host who opens the link has the identical deck rebuilt locally, and the list is never sent to the hosting server. What travels is the entry list alone; the eliminator switch, a half-finished run and the history stay on the sending screen. Over-long lists get an honest error rather than a truncated link, and a damaged link is reported as damaged. The guide Share links and privacy goes deeper.

Reading the result, and three misreadings

The draw is made the moment you press Spin; the result line under the wheel shows the full question when the rotation stops (at once under reduced motion), and the spin history keeps the most recent results for the life of the tab. Three readings of that history are wrong, and each is wrong for the same reason: the pick comes from the bytes, not from the wheel. First, "it slowed right past the travel question, so that one nearly came up" — the rotation was calculated to stop on the winner, and no other segment was ever in contention. Second, "we got the same question twice, so the wheel is broken" — in normal mode each spin is an independent draw at 4% with no memory of the last; the no-repeat setting is eliminator mode. Third, "eliminator mode makes the remaining questions less random" — it changes what remains, and every remaining question is exactly as likely as the others at every step. If you would rather check than trust, Is the wheel rigged? lists what to look for, and the browser's network tab will show that no request carries your questions.

Related wheels

To pick who answers, spin names on the custom wheel — eliminator mode ensures everyone gets exactly one turn. The Zodiac Wheel is a themed icebreaker in itself, and the Letter Wheel powers category games when questions run dry. For rosters, meetings and households, read Using the wheel for groups, and browse the rest of the guides.

Frequently asked questions

Original content, local processing: the question deck was written for this site and your edits to it never leave your browser. Draw mechanics are on the methodology page.