80–100%
Score available
Enough weighted dimensions have evidence for a reconstructible score.
SCORING MODEL / PUBLIC RUBRICS
A score measures the model. Confidence measures the evidence.
The two values never get multiplied together. Every dimension is versioned, every important claim has provenance, and unknown dimensions stay outside the score.
CAN I BUILD IT?
Vibeability
MVP scope clarity15%
Implementation complexity25%
External dependency risk15%
Non-code burden15%
Time and budget fit15%
Maintainability and reliability15%
SHOULD I BUILD IT?
Opportunity
Demand evidence25%
Commercial intent15%
Competitive whitespace20%
Distribution reachability15%
Monetization plausibility15%
Differentiation durability10%
THE COVERAGE GATE
The product earns the right to show a score.
60–79%
Provisional range
The UI shows a low-to-high range and names the missing evidence.
Below 60%
No aggregate score
The result becomes a structured validation plan, not a fake midpoint.
ANTI-CERTAINTY RULES
What the score is not allowed to do.
- Search volume alone cannot produce strong demand.
- Competitor count cannot prove whitespace.
- A price anchor cannot prove your buyer will pay.
- A fatal flag needs strong primary evidence.
- A stale source cannot support a decisive claim alone.
- A language model cannot change the final arithmetic.
YOUR NEXT MOVE
See the score. Inspect the reason.
A useful model does not hide the uncertainty that could change your decision.