fix: address PR review issues for event comparison and backtests

Critical fixes:
- Add validate_and_calculate_units() helper with proper error handling
- Handle division by zero for entry_spot in refresh_preview() and render_report()
- Add server-side validation for initial_value > 0
- Add try/except for derive_entry_spot() to handle fixture source limitations

Important improvements:
- Add dynamic default dates with get_default_backtest_dates()
- Add validate_date_range_for_symbol() for symbol-specific date bounds
- Add SYMBOL_MIN_DATES validation for backtests
- Update date_range_hint based on selected symbol

Tests:
- Add test_page_validation.py with 21 tests for:
  - validate_and_calculate_units edge cases
  - validate_date_range_for_symbol bounds checking
  - get_default_backtest_dates dynamic generation
  - SYMBOL_MIN_DATES constant verification

app/pages/event_comparison.py

@@ -13,6 +13,18 @@ from app.services.event_comparison_ui import EventComparisonPageService
 logger = logging.getLogger(__name__)
 
 
+def validate_and_calculate_units(initial_value: float, entry_spot: float) -> tuple[float, str | None]:
+    """Validate inputs and calculate underlying units.
+
+    Returns (units, error_message). If error_message is not None, units is 0.0.
+    """
+    if initial_value <= 0:
+        return 0.0, "Initial portfolio value must be positive."
+    if entry_spot <= 0:
+        return 0.0, "Cannot calculate units: entry spot is invalid. Please select a valid preset."
+    return initial_value / entry_spot, None
+
+
 def _chart_options(dates: tuple[str, ...], series: tuple[dict[str, object], ...]) -> dict:
     return {
         "tooltip": {"trigger": "axis"},
@@ -143,10 +155,13 @@ def _render_event_comparison_page(workspace_id: str | None = None) -> None:
             selected_summary.clear()
             with selected_summary:
                 ui.label("Scenario Summary").classes("text-lg font-semibold text-slate-900 dark:text-slate-100")
-                # Calculate underlying units from initial value and entry spot
-                computed_units = 0.0
-                if entry_spot is not None and entry_spot > 0:
-                    computed_units = float(initial_value_input.value or 0.0) / entry_spot
+                # Calculate underlying units with validation
+                initial_value = float(initial_value_input.value or 0.0)
+                computed_units, units_error = (
+                    validate_and_calculate_units(initial_value, entry_spot)
+                    if entry_spot is not None
+                    else (0.0, "Entry spot unavailable.")
+                )
                 with ui.grid(columns=1).classes("w-full gap-4 sm:grid-cols-2 lg:grid-cols-1 xl:grid-cols-2"):
                     cards = [
                         (
@@ -165,8 +180,11 @@ def _render_event_comparison_page(workspace_id: str | None = None) -> None:
                         ):
                             ui.label(label).classes("text-sm text-slate-500 dark:text-slate-400")
                             ui.label(value).classes("text-xl font-bold text-slate-900 dark:text-slate-100")
-                if entry_spot_error:
-                    ui.label(entry_spot_error).classes("text-sm text-amber-700 dark:text-amber-300")
+                # Show validation errors (units_error takes priority, then entry_spot_error)
+                display_error = units_error or entry_spot_error
+                if display_error:
+                    tone_class = "text-rose-600 dark:text-rose-300" if "must be positive" in display_error else "text-amber-700 dark:text-amber-300"
+                    ui.label(display_error).classes(f"text-sm {tone_class}")
 
         def render_result_state(title: str, message: str, *, tone: str = "info") -> None:
             tone_classes = {
@@ -218,8 +236,13 @@ def _render_event_comparison_page(workspace_id: str | None = None) -> None:
                     preset_slug=str(option["slug"]),
                     template_slugs=template_slugs,
                 )
-                # Calculate underlying units from initial value and entry spot
-                preview_units = preview_initial_value / preview_entry_spot if preview_entry_spot > 0 else 0.0
+                # Validate and calculate underlying units
+                preview_units, units_error = validate_and_calculate_units(preview_initial_value, preview_entry_spot)
+                if units_error:
+                    metadata_label.set_text(f"Preset: {option['label']} — {option['description']}")
+                    scenario_label.set_text(units_error)
+                    render_selected_summary(entry_spot=preview_entry_spot, entry_spot_error=units_error)
+                    return units_error
                 
                 if workspace_id and config is not None and reseed_units:
                     # Recalculate from workspace config
@@ -282,7 +305,7 @@ def _render_event_comparison_page(workspace_id: str | None = None) -> None:
             result_panel.clear()
             template_slugs = selected_template_slugs()
             try:
-                # Get initial portfolio value and calculate underlying units
+                # Get initial portfolio value and calculate underlying units with validation
                 initial_value = float(initial_value_input.value or 0.0)
                 # Get entry spot from preview
                 option = preset_lookup.get(str(preset_select.value or ""))
@@ -293,7 +316,12 @@ def _render_event_comparison_page(workspace_id: str | None = None) -> None:
                     preset_slug=str(option["slug"]),
                     template_slugs=template_slugs,
                 )
-                underlying_units = initial_value / entry_spot if entry_spot > 0 else 0.0
+                # Validate and calculate underlying units
+                underlying_units, units_error = validate_and_calculate_units(initial_value, entry_spot)
+                if units_error:
+                    validation_label.set_text(units_error)
+                    render_result_state("Input validation failed", units_error, tone="warning")
+                    return
                 
                 report = service.run_read_only_comparison(
                     preset_slug=str(preset_select.value or ""),