docs: add GLD pricing and underlying selector roadmap items

docs/roadmap/ROADMAP.yaml

@@ -1,5 +1,5 @@
 version: 1
-updated_at: 2026-03-26
+updated_at: 2026-03-27
 structure:
   backlog_dir: docs/roadmap/backlog
   in_progress_dir: docs/roadmap/in-progress
@@ -13,9 +13,13 @@ notes:
   - Pre-alpha policy: we may cut or replace old features without backward compatibility until alpha is declared.
   - Alpha migration policy: once alpha is declared, compatibility only needs to move forward; backward migrations are not required.
 priority_queue:
+  - PRICING-001
+  - PRICING-003
   - EXEC-002
+  - PRICING-002
   - DATA-002A
   - DATA-001A
+  - DATA-004
   - OPS-001
   - BT-003
   - BT-002A
@@ -39,10 +43,14 @@ recently_completed:
   - CORE-002B
 states:
   backlog:
+    - PRICING-001
+    - PRICING-003
+    - PRICING-002
+    - EXEC-002
     - DATA-002A
     - DATA-001A
+    - DATA-004
     - OPS-001
-    - EXEC-002
     - BT-003
     - BT-002A
   in_progress: []

docs/roadmap/backlog/DATA-004-underlying-instrument-selector.yaml

@@ -0,0 +1,25 @@
+id: DATA-004
+title: Underlying Instrument Selector
+status: backlog
+priority: P2
+size: L
+depends_on:
+  - PRICING-001
+acceptance_criteria:
+  - User can choose underlying instrument for options evaluation (GLD vs GC=F)
+  - Settings page exposes underlying selector with brief descriptions
+  - Workspace-scoped preference persisted
+  - Overview and Hedge pages reflect chosen underlying for price display and options strategy
+  - GC=F options data source placeholder exists (not yet wired to live data)
+  - UI clearly indicates which underlying is active and whether data is live or placeholder
+notes:
+  - GLD is the MVP supported underlying with live data via yfinance
+  - GC=F futures options will require a new data source (future roadmap item)
+  - For now, GC=F selection can show "coming soon" or use estimated/stubbed pricing
+  - Each underlying has different contract specs (100 shares vs futures contract multiplier)
+implementation_hints:
+  - Add underlying field to workspace/portfolio settings
+  - Extend app/domain/instruments.py with GC=F metadata (futures contract specs)
+  - Add underlying selector UI in settings
+  - DataService routes quote/options requests based on selected underlying
+  - For GC=F, stub with estimated pricing or clear "data source needed" state

docs/roadmap/backlog/PRICING-001-gld-expense-ratio-decay.yaml

@@ -0,0 +1,21 @@
+id: PRICING-001
+title: GLD Expense Ratio Decay Correction
+status: backlog
+priority: P0
+size: S
+depends_on: []
+acceptance_criteria:
+  - GLD weight_per_share uses the current expense-adjusted backing (~0.0919 oz/share in 2026), not the naive 0.1 oz/share
+  - Expense ratio decay formula implemented: ounces_per_share = 0.10 * e^(-0.004 * years_since_2004)
+  - All GLD conversions throughout the app use the corrected rate
+  - Domain tests prove conversion math matches research document examples
+  - Existing GLD-based tests updated to use corrected backing
+notes:
+  - See docs/GLD_BASIS_RESEARCH.md for the research basis
+  - Current hardcoded 0.1 oz/share in app/domain/instruments.py is incorrect
+  - This is a data/modeling fix, not a UI change
+  - Should be implemented before any GC=F comparison features
+implementation_hints:
+  - Update app/domain/instruments.py to compute GLD backing dynamically
+  - Add a data source for GLD NAV/ounce (can start with hardcoded 0.0919 approximation)
+  - Verify app/domain/portfolio_math.py and app/domain/backtesting_math.py use corrected conversions

docs/roadmap/backlog/PRICING-002-gld-gcf-basis-display.yaml

@@ -0,0 +1,23 @@
+id: PRICING-002
+title: GLD/GC=F Basis Display
+status: backlog
+priority: P1
+size: M
+depends_on:
+  - PRICING-001
+acceptance_criteria:
+  - Overview page shows both GLD implied spot and GC=F adjusted for comparison
+  - Basis indicator visible with threshold-based coloring (green < 25bps, yellow 25-50bps, red > 50bps)
+  - Warning displayed when basis exceeds normal bounds
+  - After-hours pricing gap noted when GLD is stale vs GC=F
+  - Display "last regular session close" vs "current futures indication" where appropriate
+notes:
+  - GLD only trades US market hours (9:30 AM - 4 PM ET)
+  - GC=F trades 23 hours/day, 6 days/week
+  - Contango is typically ~$10/oz, causing GLD to appear at "discount" vs GC=F/10
+  - During stress (backwardation), GLD can appear at "premium"
+implementation_hints:
+  - Add basis calculation to price display logic
+  - Use GLD price / GLD_OUNCES_PER_SHARE for implied spot
+  - Use GC=F - contango_estimate for fair value comparison
+  - Consider adding a small info tooltip explaining basis

docs/roadmap/backlog/PRICING-003-hedge-strike-gld-correction.yaml

@@ -0,0 +1,20 @@
+id: PRICING-003
+title: Hedging Strike Calculation Using True GLD Price
+status: backlog
+priority: P0
+size: S
+depends_on:
+  - PRICING-001
+acceptance_criteria:
+  - Hedge page strike selection uses true GLD price, not converted from GC=F
+  - Contract count calculation uses corrected GLD backing: ceil(gold_ounces / (100 * GLD_OUNCES_PER_SHARE))
+  - ATM/OTM strike calculations use GLD current price times backing ratio
+  - UI shows correct "contracts needed" for user's gold ounce exposure
+notes:
+  - Previous naive calculation assumed 10 shares per ounce
+  - Correct ratio is ~10.9 shares per ounce (1 / 0.0919) in 2026
+  - Do NOT convert from GC=F to GLD for hedging — use GLD price directly
+implementation_hints:
+  - Update app/pages/hedge.py strike calculations
+  - Update app/services/strategy_templates.py estimated contract calculations
+  - Verify app/components/strategy_panel.py displays correct contract count