docs: review backlog after backtest ui sprint

docs/ROADMAP.md

@@ -414,12 +414,66 @@ DATA-001 (Price Feed)
 
 **Dependencies:** BT-003, BT-001A
 
+### BT-001B: Backtest Defaults from Saved Portfolio [P1, S] **[depends: BT-001A, PORT-001A]**
+**As a** portfolio manager, **I want** the backtest forms to prefill from my saved portfolio settings **so that** I do not have to re-enter start value, start price, weight, and loan inputs across pages.
+
+**Acceptance Criteria:**
+- `/backtests` defaults derive from the canonical saved portfolio/settings state
+- `/event-comparison` defaults derive from the same canonical saved portfolio/settings state
+- UI makes the implied starting position explicit (start value, start price, weight)
+- Changes in settings are reflected in both backtest pages on reload
+- Browser test verifies saved settings influence backtest defaults
+
+**Technical Notes:**
+- Reuse the existing portfolio/settings persistence path
+- Avoid duplicating position-derivation logic between settings and backtest pages
+- Keep BT-001A/BT-003A read-only with respect to persistence; settings remains the source of truth
+
+**Dependencies:** BT-001A, PORT-001A
+
+### BT-003B: Event Comparison Drilldown [P1, M] **[depends: BT-003A]**
+**As a** portfolio manager, **I want** to inspect why one ranked strategy beat another **so that** the comparison page explains survival, cost, and path behavior instead of only showing a summary table.
+
+**Acceptance Criteria:**
+- Selecting a ranked strategy shows its daily path details
+- UI exposes at least:
+  - margin call days hedged/unhedged
+  - total option payoff realized
+  - total hedge cost
+  - final equity
+- UI highlights the worst LTV point and any breach dates
+- User can compare the selected strategy path against the unhedged baseline and one other ranked strategy
+- Browser test verifies drilldown content updates when selecting a ranked result
+
+**Technical Notes:**
+- Reuse existing `daily_path` and summary metrics from BT-001/BT-003 results
+- Keep this slice read-only; no saved reports required yet
+- Prefer progressive disclosure below the ranked table rather than a separate workflow
+
+**Dependencies:** BT-003A
+
+### BT-001C: Shared Historical Fixture/Test Provider Cleanup [P2, S] **[depends: BT-001A, BT-003A]**
+**As a** developer, **I want** one shared deterministic historical fixture provider for backtest UI paths **so that** seeded browser tests and demo scenarios stay consistent and do not duplicate historical-window logic.
+
+**Acceptance Criteria:**
+- Deterministic historical fixture/provider logic is centralized instead of duplicated across page-specific services
+- Supported seeded windows are explicit and fail closed outside allowed ranges
+- Both `/backtests` and `/event-comparison` use the shared deterministic provider in browser-tested demo paths
+- Tests cover the supported-window contract and failure mode
+
+**Technical Notes:**
+- Prefer a small shared test/demo provider abstraction over page-local fixture classes
+- Keep live providers separate from deterministic seeded UI/demo behavior
+
+**Dependencies:** BT-001A, BT-003A
+
 ## Implementation Priority Queue
 
-1. **BT-001A** - Put the synthetic backtest engine behind a thin product UI/read path
-2. **BT-003A** - Expose event preset comparisons in the UI
+1. **BT-001B** - Prefill backtest UIs from canonical saved portfolio settings
+2. **BT-003B** - Add event-comparison drilldown and explanation of ranking outcomes
 3. **PORT-003** - Historical LTV visibility and export groundwork
 4. **BT-002** - Upgrade backtests to real daily options snapshots
-5. **EXEC-001** - Core user workflow
-6. **EXEC-002** - Execution capability
-7. Remaining features
+5. **BT-001C** - Consolidate deterministic historical fixture/demo provider logic
+6. **EXEC-001** - Core user workflow
+7. **EXEC-002** - Execution capability
+8. Remaining features