remove min_target_probability gate + add chart time-range presets

min_target_probability is gone: it filtered on the probability model the
calibration has repeatedly shown to be weak and overconfident, it was redundant
with the momentum gate, and as an off-by-default knob it just invited bad tuning.
Removed from the backend gate, activation config/schema, the frontend mirror
(qualifiesSetup / activationSummary), and ActivationSettings. The probability
model stays where it does real work (primary-target selection + display).

Charts: with multi-year history the all-bars default was unreadable. Added
time-range presets (1M / 3M / 6M / YTD / 1Y / 3Y / 5Y / All), defaulting to 1Y;
clicking a preset always re-applies (snaps back after a manual zoom). Y-axis
autoscale and wheel-zoom / drag-pan were already there.

339 backend tests pass; frontend build clean.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

frontend/src/lib/qualification.ts

@@ -46,9 +46,6 @@ export function qualifiesSetup(setup: TradeSetup, config: ActivationConfig): boo
     return false;
   }
   if (config.exclude_conflicts && (setup.risk_level ?? '') !== 'Low') return false;
-  if (config.min_target_probability > 0 && bestTargetProbability(setup) < config.min_target_probability) {
-    return false;
-  }
   return true;
 }
 
@@ -59,6 +56,5 @@ export function activationSummary(config: ActivationConfig): string {
   parts.push(`R:R ≥ ${config.min_rr.toFixed(1)}`, `conf ≥ ${config.min_confidence.toFixed(0)}%`);
   if (config.require_high_conviction) parts.push('high-conviction');
   if (config.exclude_conflicts) parts.push('clean');
-  if (config.min_target_probability > 0) parts.push(`target ≥ ${config.min_target_probability.toFixed(0)}%`);
   return parts.join(' · ');
 }

frontend/src/lib/types.ts

@@ -162,7 +162,6 @@ export interface ActivationConfig {
   min_momentum_percentile: number;
   min_rr: number;
   min_confidence: number;
-  min_target_probability: number;
   require_high_conviction: boolean;
   exclude_conflicts: boolean;
 }