◉Krillion Pedia

2026-11-21

Krillion Scoring: How Rarity Tiers Are Decided

Krillion does not simply reward a correct answer. Every accepted answer is placed into a rarity band, and the bands decide what the answer is worth. This page explains how those bands are produced — the parts that are mechanical, the part that is human judgement, and the one detail that confuses almost everyone.

The four bands

Band What it means Shown as
Schooler The answer most players type. Valid, obvious, low value. 30 pts
Rare A real step off the obvious pick, still inside the category. 60 pts
Deep Cut The answer that makes the other players wish they had thought of it. 85 pts
One in a Krillion The gem the prompt was built around. 100 pts

Values as displayed in the game at the time of writing. The bands are relative to the prompt: a Rare answer on one prompt may be easier to find than a Schooler answer on another.

How an answer set is built

The game's own FAQ describes the process directly: answer sets are produced by aggregating datasets — dictionaries, wiki pages, encyclopedias, whatever the question calls for — and then merging and cleaning the result into a single set. That explains the shape of the answers you see. They are documented, enumerable entries, not free text. It also explains the single most common complaint: if a source was missed, a genuinely fair answer can be absent.

How rarity is decided

Rarity is not one algorithm. According to the game's FAQ, it depends on the prompt:

  • Mechanical prompts. Some can be scored entirely against a familiarity database — how often a word or name actually turns up in what people read and write. Low frequency means a deep band.
  • Judgement prompts. Others cannot be reduced to a number that way. Those get a combination of LLM assistance and a manual pass by the creator.

So "rarer" means "less common in the wild", not "harder to spell" or "more obscure to me personally".

Why the 100-point gem is sometimes less rare than a Deep Cut

This is the detail that trips people up. The gem is not chosen on rarity alone. Past a certain depth, rarity gets hard to disambiguate — several answers are all roughly equally uncommon — so the gem is also picked for fun factor: the answer that is most satisfying to land on.

The creator's own framing is that some collected answers are so obscure that maybe one player would ever type them, and a gem nobody can get is a bad gem. So the gem is deliberately kept reachable by at least a few players.

Practical consequence: if you are chasing maximum points, the gem is worth trying for, but a well-chosen Deep Cut is often the more reliable target.

Where autocorrect fits in

Many answers carry aliases, because people use different names for the same thing, and autocorrect checks aliases as well as primary names. That is generous by design — and occasionally surprising. The documented example is Threebanded Sheephead having the alias "sea donkey", which is one letter away from "sea monkey", an entirely different thing.

The takeaway is not to type carefully in a general sense; it is to type the plain, documented name, because near-misses can snap onto a different entry.

What this means for how you play

  1. Rarity is about the wild, not about you. "I have never heard of it" is not the same as "it is rare".
  2. Category first. A deep answer to the wrong category does not exist.
  3. The bracket is a filter, not a hint. Failing a stated rule removes the answer entirely.
  4. One band lower, reliably, beats one band higher, occasionally.

Common questions

Is the scoring the same in every mode? The four bands are the game's scoring language. Modes and packs can vary in timing and format, so treat the daily dive as the reference.

Can I see an answer's rarity before submitting? No. Rarity is only revealed once an answer is accepted, which is why practice modes matter.

Why do two answers that feel equally rare land in different bands? Because the band is computed from frequency data or a manual pass, not from how obscure the answer feels to one player.

Keep reading