https://wheelcue.com/research/ Original studies of how random picks really behave in classrooms, giveaways and games, built on simulations of real draws. en-us Thu, 22 Oct 2026 12:00:00 GMT https://wheelcue.com/assets/img/icon-192.png https://wheelcue.com/research/ https://wheelcue.com/research/random-partners-week-after-week/ https://wheelcue.com/research/random-partners-week-after-week/ Thu, 22 Oct 2026 12:00:00 GMT Pair a class at random every week and last week's partners keep turning up together again. In a class of 24, 40.6% of weeks bring back at least one pair from the week before, and by week 8, 74.8% of students have had the same partner twice. Pressing Generate again can rule out last week's pairs in a press or two, but ruling out every earlier partner gets impractical within a term. We ran 3,682,151 simulated presses of Wheel Cue's team generator, with the same random.js and splitting steps, worked out the exact odds, and built a planner that gives a class of 24 new partners for 23 weeks while keeping the pairs random. https://wheelcue.com/research/how-many-houses-for-every-halloween-candy/ https://wheelcue.com/research/how-many-houses-for-every-halloween-candy/ Mon, 19 Oct 2026 12:00:00 GMT If every house hands out one random piece from the same bag, how many houses until a trick-or-treater has every kind? With the pretend 20-piece bag on Wheel Cue's Halloween wheel, 8 kinds take 34.6 houses on average, and a full set within 30 houses comes only about half the time. We worked out the exact odds, checked them against 89.6 million simulated spins through the site's real random code, and turned them into a class activity that needs no real candy. https://wheelcue.com/research/raffle-odds-what-more-tickets-buy/ https://wheelcue.com/research/raffle-odds-what-more-tickets-buy/ Thu, 15 Oct 2026 12:00:00 GMT Once there are several prizes, twice the tickets can fall well short of twice the chance of winning something; a ticket that makes the raffle bigger is worth a little less than a full share; and in a giveaway where each person can win once, the order of the draw changes who is likely to get the top prize. We ran 41.5 million simulated spins through the code behind every Wheel Cue pick and checked them against the exact odds. https://wheelcue.com/research/are-random-teams-balanced/ https://wheelcue.com/research/are-random-teams-balanced/ Mon, 12 Oct 2026 12:00:00 GMT A random split is fair to every student, but it is not built to be even. In a class of 24 rated 1 to 5, four random teams leave the strongest team's average a full point or more above the weakest's in 58.2% of splits, and the four strongest students land on four different teams only 12.2% of the time. We ran 3,974,246 simulated presses of Wheel Cue's team generator through its real code, worked out the exact odds alongside, and measured which simple fixes actually work. https://wheelcue.com/research/why-random-feels-rigged/ https://wheelcue.com/research/why-random-feels-rigged/ Wed, 07 Oct 2026 12:00:00 GMT A wheel that lands on the same name twice is not broken. On a 25-name wheel, the chance that some name has come up twice passes 50% at pick 7, and in 100 fair coin flips the most likely longest streak is 6. We ran about 46 million simulated spins through the code behind every Wheel Cue pick, worked out the exact math alongside, and measured how often fair randomness repeats, streaks and clumps. https://wheelcue.com/research/how-many-spins-until-everyone-gets-a-turn/ https://wheelcue.com/research/how-many-spins-until-everyone-gets-a-turn/ Sat, 03 Oct 2026 12:00:00 GMT If the winner stays on the wheel, a class of 25 needs 95.4 spins on average before every student has been called at least once, and 1 class in 10 needs 135 or more. Here is why, what it means for a lesson and a week of random calling, and the routine that shares turns exactly.