Day 23: The anti-signal — when 6/6 is wrong

Some orgs game every signal. Building your false-positive list is week 4 work.

Yesterday

Yesterday you set alerts. Today: the override layer that catches the orgs where the alerts will lie to you.

Why this signal matters

Frameworks are gameable, especially public ones. A devtools agency, a YC alum farming OSS contributions, a team running a single-customer consultancy that looks like a startup — all can score 6/6 on the public composite. Your false-positive list is the personal antibody.

The 5-minute exercise

1List 3 archetypes you've seen game public-data signals (agencies, single-customer consultancies, OSS-funded research labs).
2For each, note the signal that betrays it on closer reading (e.g. agency → look at customer-named repos).
3Add a manual 'sanity check' step to your Monday rhythm: top 1 from watchlist gets the 5-minute archetype check.
4Save the archetype list. Update quarterly.

What you’re filtering for

A written, dated false-positive list with at least 3 archetypes and the manual check that catches each. The list is private — it's competitive advantage, not content.

Edge case

A single high-trust founder you've already met can override a low score the same way an archetype can override a high score. Both are personal-judgment overrides on top of the framework, not replacements.

Bonus

The archetype list is the most personal artifact in this whole challenge. It encodes your specific bias-correction. Two investors with the same 6-signal framework will produce different deal flow because their archetype lists differ.

Tomorrow

Tomorrow: making the score legible to a co-investor — the share template that turns a number into a paragraph.

Curriculum: /challenge · Methodology: /methodology · Paper: ssrn.com/abstract=6606558