# Study 000B Methods

## Design

Primary flagship companion study conducted after `Study 000A` to examine whether later specialization coincided with adaptive efficiency, persistent unexplained HR burden, or both.

## Relationship to Study 000A

`Study 000A` established that the system adapted through selective flexibility and specialization.

`Study 000B` does not reopen that question. It stands alongside the flagship and asks a second primary program question:

- did external output become more efficient relative to cardiovascular cost?
- did some background recovery markers move in a favorable direction as specialization increased?
- did session-level internal burden disappear, or did part of it persist?

## Core question

As running specialization increased in the later high-resolution window, did the system show better efficiency, persistent unexplained HR burden, or both together?

## Packaged source tables

| table_name | relative_path | bytes | sha256 |
|---|---|---|---|
| v7_pillar_conserved_mechanics_runs_v1.csv | source_tables/v7_pillar_conserved_mechanics_runs_v1.csv | 204809 | baaba0ea64e74da234a19e8dd1eaac864ead0163ef5c52ea505e8bdf92c68f0c |
| v7_total_fitness_profile_timeline_v3.csv | source_tables/v7_total_fitness_profile_timeline_v3.csv | 618684 | e9a1c71bcf06f880c0b7fd1c011336ee3323687e4fd8d5cb857508e727c06028 |
| v7_food_pdf_garmin_overlap_v1.csv | source_tables/v7_food_pdf_garmin_overlap_v1.csv | 15436 | 82fc9280f4e7d36865086c1cd02b898b081ca618c4db37a7e39c8092a393965a |

## Analytic structure

1. QC-pass high-resolution runs were linked back to the unified daily timeline.
2. Run-level efficiency metrics were derived as `speed_per_hr` and `power_per_hr`.
3. Early and late 30-run windows were compared on output, cost, specialization, and next-day recovery. These windows were matched by run count rather than elapsed time, spanning `93` calendar days early and `43` calendar days late.
4. A simple descriptive HR model (`HR ~ speed + power`) was fit across QC-pass runs to estimate unexplained session HR burden after basic external-output adjustment.
5. Monthly summaries and specialization correlations were generated to characterize broader late-window trends.
6. The food overlap layer was retained only as partial exploratory late-window context.

## Interpretation boundary

This study is descriptive and within-subject. It does not prove physiology mechanism or clinical causation. It asks whether the data support a mixed pattern of better efficiency together with persistent unexplained HR burden.