Day 26: Custom composite weights

Default is 1×6. Your beat probably needs 2× on one signal and 0.5× on another.

Yesterday

Yesterday: MCP integration. Today: the personal calibration that the integration alone can't give you.

Why this signal matters

The unweighted composite (each signal +1) is the right starting point because it's calibration-light. After 26 days you have enough personal data to see which signal matters most for your beat, and a re-weighted composite outperforms the default for your specific deal flow.

The 5-minute exercise

1Pull the watchlist CSV and the 4-org scorecard from Day 14.
2For each signal: which one would you have weighted higher to better predict your hits?
3Propose new weights (e.g. commit velocity 1.5×, dependents 1.5×, others 1×).
4Recompute the composites with new weights. Compare ranking shifts.

What you’re filtering for

A reweighted composite that produces a ranking you trust more than the default for your beat. If the reweighting doesn't change the ranking meaningfully, the default is correct for you.

Edge case

Don't over-weight a signal because you remember one big hit on it. The right weighting reflects pattern across 10+ orgs, not one anchor case. Sample size discipline matters here.

Bonus

The custom weights become a competitive moat. Two analysts with the same framework but different weights produce different deal flow — and yours is calibrated to your specific outperformance pattern.

Tomorrow

Tomorrow: the 30-day retrospective. Pull out the artifacts. What did the framework actually catch?

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