All new games—Classic, Daily, Practice and Guided demo—use proportional scoring and per-answer platforms: seven questions and a maximum of 700 points. Timed rounds allow up to 15 seconds; an accepted submission can reveal early. Its answer counts are synthetic demonstration data, not survey results or PlayDaily records. A more detailed formula does not make the underlying data more accurate.
How the 1–100 score works
Points = round(answer count ÷ highest answer count × 100), with a minimum of 1 for a supported answer with a positive count. Equal counts earn equal points and share a rank. Rank alone does not determine the proportional score.
If A has 10% of all responses and is the largest group, it earns 100. B at 5% earns 50—not a fixed rank-based tier. These percentages must come from the same question, population and counting method. This example explains a formula; it is not a measured survey.
INTERACTIVE EXPLANATION · NOT A SURVEY
See how a count becomes a score.
Change the two hypothetical counts. This calculator does not change your game or submit an answer.
Use positive whole counts; the most popular answer cannot have fewer responses. Calculator limit: 1,000,000.
5 ÷ 10 × 100, rounded to the nearest whole point; a supported positive count has a minimum of 1.
A real example from our simulated bank
The kitchen test question currently has these artificial counts. The total weight across its 12 answers is 505 and the largest weight is 100. These four rows are an excerpt; the full inventory includes all 12.
| Answer | Demo count | Share of 505 | Points |
|---|---|---|---|
| fridge | 100 | 19.80% | 100 |
| oven | 83 | 16.44% | 83 |
| sink | 71 | 14.06% | 71 |
| spoon | 62 | 12.28% | 62 |
A score of 100 means “equal to the largest recorded group,” not “everyone answered this.” The expanded demo includes positive weights from 1 to 100. Scores are calculated independently; they are not rounded into visual tiers.
What the animated platform means
Ten animated platform presets represent relative popularity: level = round(answer weight ÷ largest weight × 10), clamped to 1–10 for positive weights. Level 1 shows one figure; level 10 shows 24 decorative figures, including you in the center. These are visual density levels, not actual respondent counts. A score of 76 and a score of 84 can share level 8 without earning the same points.
Supported answers, language and missing data
The current preview has 21 test questions, 244 standard answers and 342 explicit aliases across everyday life, food and world topics. English canonical answers and listed Chinese equivalents are supported. Case, surrounding spaces and Unicode forms are normalized; arbitrary translations, pinyin and typo correction are not supported. AI judging is not connected.
Normal submission of an unlisted answer leaves time to retry. Skip wait submits the current nonempty answer and ends the remaining wait: a listed answer scores normally; an unlisted answer is retained as not scored. It adds no points and has no invented crowd. This is not a judgment that the answer is wrong. A blank timeout also adds no points; the minimum of 1 only applies to accepted positive-count answers.
Do not compare a limited test-table share with population preferences. Search result totals and search ranking are not response counts. Check answer coverage and troubleshooting or inspect the current data inventory.
One scoring model across all modes
Classic, Daily and Practice now use the same relative-count formula and animated per-answer platforms. Guided demo adds visible answer hints; it is not a different scoring system. Practice is untimed; timed modes use seven 15-second rounds. Old saved runs keep their historical scores and must be restarted to use the new rules. Old-rule challenge links must be recreated. Daily remains this edition’s own UTC selection, not PlayDaily’s puzzle. Its shuffled 21-question pool rotates seven questions per day: three consecutive dates cover all 21 before reuse. Classic, Practice and Guided demo prioritize unseen questions in this browser, then least recently seen. Each single-theme pool has only seven questions and therefore repeats each run. Clearing browser storage resets this preference; it is not anti-cheat protection.
What protects a result?
The server owns the deadline, answer mapping and score. A run keeps its question version; repeated delivery of the same submission cannot add points twice. An accepted answer locks. A scoring-service result inconsistent with the frozen question must void the round rather than silently award a plausible-looking score. None of these checks makes a synthetic dataset representative.
For a different implementation, read PlayDaily’s own rules. Our proportional preview is an independent choice, not its official formula.
