Glossary term
Abstain
Definition
A node with nothing honest to say stays quiet. Zero news articles is not a mood of 50. Thin tallies do not become secret leans.
How it works
A node abstains when it has nothing honest to say, usually because its input is missing: a news node with zero articles, a fundamentals node with no filing, a cycle study over a sample too short to carry one. An abstaining node is removed from the calculation entirely, which is different from reporting mixed. A mixed node stays in the denominator and dilutes the score.
The rule has to live in the pipeline, not in a prompt. A model asked to summarize the mood of an empty set will often produce a measured paragraph anyway, because producing measured paragraphs is what it does. Enforcing no input, no vote in code prevents the tally from absorbing an invented mood.
Worked example
Two nodes lean constructive on a hypothetical name and eight abstain. The score is (2 − 0) / 2 = 1.00, the maximum the scale can print, from two opinions.
Now suppose the same eight report mixed instead. The score is (2 − 0) / 10 = 0.20, which is below +0.25 and prints mixed. The two constructive opinions did not change at all; the denominator went from 2 to 10. The active count printed next to the score is the only warning that the first reading rests on two voices.
Hypothetical, for education only. Not a recommendation and not a prediction.
Common mistake
The mistake is treating silence as a neutral opinion, for instance letting zero news articles count as a mood of 50 out of 100. Abstaining and being neutral are different acts, and the arithmetic treats them very differently. A related error is reading a thin tally as a secret strong lean because the score looks extreme. A quiet node is a legitimate output, and a tally built from very few voices should not become a headline.