How Weighted Random Selection Works
Weights let you deliberately give some entries a larger chance than others. They are useful only when the group understands why those chances differ.
Think in shares, not guaranteed outcomes
Imagine a task wheel with Review, Practice and Break. Give them weights of 3, 2 and 1. The total weight is six, so Review has three-sixths of the chance, Practice has two-sixths and Break has one-sixth. Those are one-half, one-third and one-sixth. A weight of three does not mean an entry must appear three times in the next six spins.
Multiplying all weights by the same positive factor leaves their proportions unchanged. Weights 30, 20 and 10 describe the same chances as 3, 2 and 1. You do not need to type percentages that add up to one hundred. Smaller readable numbers usually make it easier to check a setup with other people.
Enable and review the actual list
Orbit's Wheel and custom Roulette support positive entry weights. Enable weighted selection before expecting those values to affect the result. When weights are disabled, entries have equal chances regardless of the saved weight fields. Coin Flip uses its fixed two outcomes, and the fixed European Roulette preset uses its fixed outcomes.
Look for duplicate entries before checking the total. If Review appears twice, its chances combine across both entries. Duplicate display names can make a result look inconsistent even when the underlying entries were sampled correctly. For a one-person-one-chance name list, keep one entry per person and equal weights.
Orbit's stored weights use bounded precision, and selection converts accepted values into integer units. Use simple positive weights that the interface accepts. Extremely fine distinctions smaller than the supported precision will not produce a useful extra level of control. Check the displayed values after editing rather than assuming every decimal you typed was preserved.
Selection is separate from motion
The selection uses browser Web Crypto and a bounded integer sampler. Draws that would bias a remainder operation are rejected. Orbit chooses an entry from the cumulative weight shares before the wheel or Roulette animation begins. The display then moves to the already selected outcome.
A long spin, a different final landing position or a reduced-motion reveal does not change which weight share was selected. The wheel can stop at different safe positions within the winning segment. This makes repeated results less visually identical without changing their probability.
Independent draws may form streaks. If Review has a one-half chance, getting Review twice does not make Break due on the next spin. Removing entries after a result changes the list, however, and therefore changes the chances of later draws. Recalculate using only the remaining entries when you use removal.
Choose a fair rule for the task
Weighted selection is not automatically fairer than equal selection. For a classroom turn, equal chances may fit the goal. For allocating a practice activity, you might deliberately give a difficult topic more weight. Explain that rule first and obtain agreement where appropriate. Orbit cannot decide whether your reason is justified.
For a giveaway, follow its published eligibility and entry rules. If the rules allow different numbers of entries per participant, make the mapping explicit and check it before drawing. A wheel cannot verify those rules, judge legal requirements or prove that a list was complete. Do not retroactively change weights to justify an unwanted result.
Try a small example before using a long list. A three-entry setup makes it easier to check total weight and expected proportions. History can help inspect your own workflow, but a short sample will not match exact theoretical percentages. Use it as a record of individual draws, not as evidence that a random streak is a fault.