Krillion Scoring & Rarity Tiers
Facts decide whether an answer belongs. Our curated practice ratings decide its points.
Our six practice tiers
| Tier | Points | What it means here |
|---|---|---|
| Common | 10 | Familiar starting picks. |
| Familiar | 15 | A second accessible choice. |
| Uncommon | 30 | A step beyond our starting picks. |
| Rare | 60 | A less familiar choice in our practice list. |
| Deep Cut | 85 | A more specific discovery. |
| Standout | 100 | Our selected top-scoring answer. |
These labels are our judgment, not measured rarity. Familiarity differs between players and cultures. A score is fixed for this version of the question bank and does not change when another player chooses the same answer.
Official scoring is separate
The official FAQ describes a mix of familiarity data, model assistance and manual curation, depending on the prompt. It also explains that the highest-tier answer is not chosen on obscurity alone. Our points do not reproduce its algorithm and never transfer to an official account.
Names, aliases and omissions
Capitalization, leading or trailing spaces and repeated spaces do not affect matching. Listed spellings and abbreviations share a score. We do not automatically accept fuzzy matches: similar spelling can identify a different thing. A missing answer is labeled “Not in this practice list,” and you can retry within the timer.
Official Krillion tiers
The original game uses the names Plankton, Too Clever, Schooler, Rare, Deep Cut and One in a Krillion, corresponding to 10, 15, 30, 60, 85 and 100 points. “Too Clever” means the obvious obscure pick, according to the official FAQ. We use the same six numeric steps for a familiar format, but different editorial labels and independently assigned answers. A 100-point practice answer need not receive 100 points at krillion.io.
Points and dive depth
Seven-question dives can reach 700 points. Unlimited has ten questions and a 1,000-point ceiling. Fixed-length dives calculate depth as total points divided by the maximum, multiplied by 7,000 metres. For example, 600 of 1,000 points displays 4,200 m.
Infinite uses average points per completed question for depth. Its personal best is recorded only after at least seven questions, so a one-question 100 does not outrank a full session. Timed, untimed and topic records are separate.
Practice is not live statistical rarity
The scores do not change when someone else chooses an answer. They are not popularity surveys, measured selection rates or an imitation of the official familiarity algorithm. Some bounded lists are small; familiarity varies between cultures and players. Treat the ratings as a challenge within this bank, not an objective ranking of your knowledge.
Solo records stay on this browser, not in a global leaderboard. Friend matches compare server-validated scores on the same route and use total answer time for ties. Read the full rules or play a round.