OVERSCANUI KIT
no button at all a worn bezel with weight channels do not converge zero depth instrumented, full of smoke a field of real characters a camera looking out additive, so no shadow hairlines, no new shader no raster, so no scanlines

A finding's certainty in calibrated words, and only the words its evidence has earned. Every finding is judged on evidence and agreement first, marked on the grid at the left. With limited evidence and low agreement those two terms are all it shows: a confidence or a likelihood handed in is withheld and says why. A likelihood is shown only at high or very high confidence. Confidence is five steps and never a percentage. A calibrated term is a range, so "likely" is drawn from 66% to 100%, and a bare number is never turned into a word. "About as likely as not" is not a way of saying "we don't know". The rules are the IPCC's, from its 2010 guidance note for authors.

demo/likelihood.html, from this page
import 'overscan'
import 'overscan/overscan.css'

<ov-likelihood
  src="models/likelihood-seawall.json"
  label="FINDING 2.1"
></ov-likelihood>
earned: high confidence, a calibrated term
medium confidence: the likelihood is withheld
limited, low: only the summary terms
a range, so no term
low confidence, kept for a reason
variation: src="models/likelihood-pump.json" label="FINDING 4.3"
import 'overscan'
import 'overscan/overscan.css'

<ov-likelihood
  src="models/likelihood-pump.json"
  label="FINDING 4.3"
></ov-likelihood>
terminal
variation: src="models/likelihood-anomaly.json" label="FINDING 6.0"
import 'overscan'
import 'overscan/overscan.css'

<ov-likelihood
  src="models/likelihood-anomaly.json"
  label="FINDING 6.0"
></ov-likelihood>
industrial
variation: src="models/likelihood-flood.json" label="FINDING 1.4"
import 'overscan'
import 'overscan/overscan.css'

<ov-likelihood
  src="models/likelihood-flood.json"
  label="FINDING 1.4"
></ov-likelihood>
cyber
antiseptic
esper
neo
machina
holo
aegis
vector