Confidence scale (like priority scale)

December 16, 2025

We introduced a “confidence scale”, a new form of communication shorthand at my company several months ago. It’s held up well. Teammembers have conveyed useful epistimic information more regularly in meetings, messages, and memos.

Confidence scale.

Use a numeric scale—c0, c1, c2, c3, c4, c5—as shorthand to communicate confidence in a statement. This is analogous to p0 - p3 as shorthand for priority. Same as priority, c0 is highest conviction / p0 is most urgent.

[c0] a factual statement that cannot be meaningfully contested. E.g., The Hubble Space Telescope was launched on April 24, 1990 [c0]. In hindsight, these claims should be wrong <0.1% of the time.
[c1] a high confidence statement that is backed by strong research. E.g., Viasat is the largest provider of in-flight WiFi by number of installed aircraft worldwide as of 2025 [c1]. In hindsight, these claims should be wrong <1% of the time.
[c2] a medium-high confidence statement that is backed by reasonable logic and research. E.g., Reausable launch vehicles will account for the majority of orbital launches by 2030. In hindsight, these claims should be wrong <10% of the time.
[c3] a medium confidence statement that is backed by reasonable logic and minimal research. E.g., SpaceX will launch more payload mass to orbit than all other providers combined in 2026. In hindsight, these claims should be wrong <30% of the time.
[c4] a low confidence statement that is backed by loosely recalled facts or anecdote. E.g., NASA’s Artemis program will land humans on the Moon before China’s crewed lunar program does*. In hindsight, these claims should be wrong <66% of the time.
[c5] a baseless claim, often an optimistic belief. E.g., SETI will identify extraterrestrial intelligent life by 2035 [c5].

Why?

The confidence scale helps communicate epistemics concisely [c1] (fun fact: when the SOP was first released interally, we rated this previous statement a “c4”; now that we have more evidence it’s a c1). Since it reduces the cludge that comes from English language—no one says “I’m high conviction on this”, instead they say “I think …” which is imprecise—, it’ll be used more regularly and better inform our decision making.

*Note that the c4 example is forward looking; often in practice we find that “c4” is also used with statements about past events that the speaker doesn’t recall well.

Adoption is still not at 100%, but it’s on par with our team’s adoption of priority scale shorthand [c2]. We’ll 2x our per capita usage within 6 months [c2.5].