Overview

The six judges

The lenses#

Judge Reads
J-P1 Problem The pain, the user, the urgency, and the alternatives a deck claims to beat
J-P2 Solution Logic Product logic, differentiation, and how coherently the solution holds together
J-P3 Business Value / Market The market, the value, and how the business intends to make money
J-P4 Pitch Quality Clarity, narrative, structure and delivery
J-P5 Team Readiness Founder-market fit, skills, and the ability to execute
J-P6 Feasibility Roadmap, resources, and operational realism

Not every judge influences every score#

A lens contributes to a dimension at one of four levels: primary (1.00) drives the score, secondary (0.50) adds important support, advisory (0.25) provides context, and not scored means no influence at all.

Judge P1 Problem P2 Solution P3 Market P4 GTM P5 Team P6 Feasibility
J-P1 Problem primary advisory advisory advisory
J-P2 Solution Logic secondary primary advisory advisory secondary
J-P3 Business Value / Market advisory advisory primary primary advisory
J-P4 Pitch Quality advisory advisory advisory advisory advisory advisory
J-P5 Team Readiness advisory advisory primary secondary
J-P6 Feasibility advisory secondary secondary secondary secondary primary

Two things are worth reading off this table.

Pitch Quality is advisory everywhere. It is visible in the report and it never drives a dimension. This is the structural answer to "does a polished deck win here?" — presentation quality matters, but it cannot outrank weak evidence on problem, market, team or feasibility.

Team Readiness does not score Problem or Solution. A lens only scores where it has a legitimate read. That is what keeps a strong impression of the founders from bleeding into every dimension.

Why independence, structurally#

Each judge evaluates in an isolated context. Three consequences:

  • No anchoring. A lens cannot converge on another's number, so agreement between them is evidence rather than an artifact.
  • Halo effect is split. Dimensions are read by separate judges, so one strong impression cannot carry the whole scorecard.
  • Containment. An instruction hidden in a deck that reaches one context cannot reach another. See Prompt-injection safety.

The other bias controls#

Risk Control
Halo effect Dimensions split across separate judges
Generic scoring Dimension-specific prompts and criteria per judge
Overweighting presentation Pitch Quality visible, never dominant
Hidden disagreement Spread plus the judge contribution matrix
AI overreach The AI Total Score is advisory; the human decides
Assumption-filling Missing evidence becomes a gap or a question, never a guess

Where the method comes from#

The Pitch Competition dimension matrix combines three established startup-evaluation lenses rather than being assembled from prompt tricks:

  • Lean Startup — hypothesis and problem-solution logic, feeding P1 and P2.
  • Customer Development — customer, pain and validation evidence, feeding P1 and P2.
  • VC Due Diligence — market, business model, team and feasibility, feeding P3, P5 and P6.

It is thesis-first by design: a polished deck should not score high if the problem is vague, the customer unclear and the business logic thin.

The hackathon panel#

In Hackathon mode the panel is five reviewer roles — Innovation, Technical Execution, Business Value, Pitch Quality and Feasibility — reading every submission across an execution-weighted rubric. See Hackathons.

Next steps#

Updated

Was this page helpful?