You are judging a pairwise comparison between two research ideas

brainstormed for the same topic. Pick which is more promising as the

PublishedJun 8, 2026

Loading actions...

5 minBeginnerprompt21 files

Skill content

Main instructions and any bundled files for this skill.

markdown

Additional Files (20)

You are judging a pairwise comparison between two research ideas brainstormed for the same topic. Pick which is more promising as the basis for a research quest that will run end-to-end (literature → design → experiment → analysis → paper).

Topic

$topic

Clarify answers

$clarify_block

Idea A

$idea_a

Idea B

$idea_b

Output (single JSON object, no prose outside it)

{
  "winner": "A" or "B",
  "reason": "&#x3C;one short sentence explaining the decisive factor>",
  "margin": "decisive" or "narrow"
}

Judging criteria (weighted, in priority order)

Empirical tractability — given the user's budget and the measurement plan implied by the idea, can the experiment actually run in the budgeted compute? An idea that's beautiful but demands GPU-weeks is dominated by a slightly less novel idea that completes in an hour.
Novelty over the chosen comparative baseline — does the idea actually move past the baseline named in comparative_baseline? Re-deriving an established result with a new framing is worse than a small but genuine delta.
Methodological cleanness — does the idea afford a clean independent / dependent variable split? Ideas with too many confounders are dominated by tighter framings.
Falsifiability — picks a hypothesis that COULD be wrong. Vague "explore the relationship between X and Y" framings are dominated by sharp "X causes Y by mechanism M, measured by Z" framings.

Constraints

Pick exactly one winner. Ties are not allowed — break with criterion 1.
reason must cite which criterion was decisive.
margin: "decisive" if criterion 1, 2, OR 3 is unambiguous. margin: "narrow" when the two ideas only differ on falsifiability or on a soft criterion. The aggregator uses margin to break tournament-level ties; narrow wins still count as one vote.
No prose outside the JSON object. No code fences. No commentary.

View Original Source

Related Skills

General

PromptBeginner5 minmarkdown

Untitled Skill

193

Jan 12, 2026

General

PromptBeginner5 minmarkdown

Frontend Typescript Linting.mdc

TypeScript and ESLint rules that MUST be followed when creating, modifying, or reviewing any file under apps/frontend/, including .ts, .tsx, .js, and .jsx files. Also apply when discussing frontend li...

160

Feb 15, 2026

General

PromptBeginner5 minmarkdown

2. Apply Deepthink Protocol (reason about dependencies

risks

127

Jan 15, 2026