A weight is a relative chance
For a random draw with fixed eligible weights of 1, 1 and 2, the third entry has 2 out of 4 of the total weight: 50%. The first two each have 25%. Those percentages describe a draw, not a promise about a short sequence.
Removing or locking an entry changes the eligible pool. Adaptive can also change effective weights between draws. Read the current state rather than relying on the original spreadsheet alone.
Pick the behavior you need
| Preset | What it does |
|---|---|
| Pure | Keeps base weights unchanged after a pick. Custom weights still matter. |
| Balanced / Strong fairness | Reduce a selected entry’s weight and let other entries recover over later rounds; Strong applies stronger adjustments. |
| No repeat | Temporarily makes the most recent pick ineligible for the next draw. It can return after another pick. |
| Cooldown | Temporarily reduces a selected entry’s effective weight; it is not an activity-long ban. |
| Pity | Adds a growing boost to entries marked rare while they wait to be selected. |
Watch the rule work
- Open Adaptive and inspect the list in Names.
- Choose a preset from the available controls or Advanced.
- Spin once, then compare the visible recent picks and weights.
- Open History to inspect or undo a previous result. If you change the roster or rules, review the current state again.
Explain it before the round
Use plain language: “We are reducing repeat picks across questions” or “Each ticket has the weight shown.” Do not describe adaptive odds as fixed equal odds. For a one-award-per-entry event, use the separate hosted rule described in Avoid repeat winners.
ONE MORE THING
Common questions
Does Pure mean every name has the same chance?
Only if all eligible entries have equal base weights. Pure keeps those weights unchanged between draws. Duplicate entries and custom weights can give a displayed name more total chance.
Link to this answerDoes fairness guarantee everyone wins equally?
No. Balanced and Strong fairness adjust future weights; they do not guarantee a fixed distribution over a short session. Use visible weights and history to explain what happened.
Link to this answer