← Results index

Separate meta strategies from individual strategies

2026-08-17T21:04:09Z Website/UI completed Split the mixed Strategy catalog into dedicated Individual Strategies (docs/catalog.html, 139 rows) and Meta Strategies (new docs/meta-strategies.html, 13 walk-forward + 2 continuous-weight rows), added primary navigation Strategies → Meta Strategies → Experiments → Results and secondary sub-nav on both pages, updated all internal links at their producers, preserved every meta metric/plot/allocation/description/link, and verified with static bundle, pipeline, and browser/link checks.
catalog meta-strategies navigation feature

Objective

Separate the current mixed strategy presentation (Strategy catalog table plus Meta strategy reports and Continuous-weight tables all on catalog.html) into dedicated individual- and meta-strategy surfaces. Create a dedicated Meta Strategies page, leave individual strategies on the existing Strategies page, remove meta strategies from that catalog, add clear navigation among Individual Strategies, Meta Strategies, Experiments, and Results/Changelog, preserve all existing meta-strategy metrics/plots/allocations/descriptions/links during the move, update affected internal links at their producer/source rather than hand-patching generated output, and confirm no internal link regressions with automated browser/link checks.

Approach

Added Meta Strategies to the locked primary navigation in research/reports/nav.py (Home → Strategies → Meta Strategies → Experiments → Signals → Plan → Progress → Coverage → Results → Pipeline) and updated both header renderers. In research/reports/scaffold.py split the monolithic render_index into individual-only (render_index now renders Strategy catalog — Individual Strategies with 139 rows, lede linking to Meta Strategies, sub-nav Individual|Meta|Experiments|Results, and mounts only the strategy catalog) and dedicated render_meta_strategies_index (H1 Meta Strategies — Walk-Forward Baskets & Continuous-Weight, lede back to Individual, same sub-nav, meta-row badges, catalog-controls with search/source/leverage, and verbatim tables #meta-strategy-catalog (13 rows from META_REPORTS) and #weighted-meta-catalog (2 rows) with identical column sets and leverage-filter swap via catalog-equalize). Updated scaffold_site to write both docs/catalog.html and docs/meta-strategies.html, plus refresh_meta_catalog_artifacts, and render_home to point Capabilities/Explore and lede to meta-strategies.html instead of catalog.html#meta-strategy-catalog. Changed research/reports/home_publish.py links.meta_catalog to meta-strategies.html and added meta-strategies.html to research/scripts/check_static_reports.py REQUIRED. Regenerated docs via python3 -m research.reports scaffold (deterministic payloads, no DB mutation). Updated tests (test_home_dashboard, test_home_dashboard_browser, test_advanced_web_browser) to expect the new split and targeting meta-strategies.html for meta-specific assertions.

Files changed

Validation

Results

Individual Strategies now live solely at docs/catalog.html (Strategies in nav, Strategy catalog — Individual Strategies, 139 AS/non-AS rows with search/source/leverage/equalize, badge counts 139, no meta tables). Meta Strategies live at new docs/meta-strategies.html (Meta Strategies in nav, Meta Strategies — Walk-Forward Baskets & Continuous-Weight, 13 walk-forward rows + 2 continuous-weight rows, same columns/badge/leverage behavior, leverage cohorts <em>-lev-</em>.html still reachable, Custom Meta builder linked). Primary navigation is Home → Strategies → Meta Strategies → Experiments → Signals → Plan → Progress → Coverage → Results → Pipeline on every shell, with a secondary pill nav Individual Strategies · Meta Strategies · Experiments · Results/Changelog on both catalog and meta pages and Home Capabilities/Explore updated. All meta detail pages, JSON payloads, QuantStats tearsheets, and custom-meta builder remain at identical URLs with unchanged metrics/plots/allocations. Every internal link was updated at its producer, not hand-patched, and automated checks confirm zero stale anchors and no link regressions.

Links

Source artifact: research/results/qm-s0t.md