Creative evaluation: splitting novelty into R1 and R2
Treating originality as one score breaks the moment a second round exists. Splitting it into R1 (independent novelty) and R2 (incremental value over the existing set) fixes the metric.
TL;DR — One "originality" score silently assumes every idea is judged against nothing. That is only true in round one. Split it: R1 = independent novelty (novel given no peers), R2 = incremental novelty (how much it adds over the existing set).
The bug the split fixes
Originality is relative — but relative to what? In a blind first round, the fair baseline is "given no other ideas, is this new?" By round two, participants have read everyone else's work, so the honest question changes: "given everything already on the table, does this add something?"
Scoring both with the same number is the metric equivalent of measuring height at two different scales.
R1 — independent novelty
- Judged blind, against the field as a whole.
- Question: how distinct is this from what the other participants independently produced?
- High R1 means the idea is unlikely to have been reached by the group's default path.
R2 — incremental novelty
- Judged openly, against the accumulated set.
- Question: how much does this advance, combine or overturn what already exists?
- A build-on can score high here even if its R1 is low — that is the point.
Why it matters in practice
Without the split, round two looks like it "produces less original work", and teams wrongly conclude the second round is a waste. With the split, round two has its own metric and its own value: combination and deepening, not raw novelty.
Implementation note
The two scores are separate dimensions with separate rules for who may assign them (the author never scores their own originality). See the P-metrics for how they sit alongside the other dimensions.
Correcting a metric after you see it fail in production is not a bug report — it is how the rubric is supposed to evolve.