backtest: add min target-probability sweep

Re-applies the activation gate at several min_target_probability thresholds
(60→30, other conditions fixed) over the already-replayed candidates, so the
trade-off between how many setups qualify and their expectancy is visible in one
table — the cheap "optimize" half of Phase 2. Candidates now carry meets_core +
best_prob so the sweep needs no re-replay. New sweep table in BacktestPanel with
the current threshold starred.

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

app/services/backtest_service.py

@@ -160,6 +160,12 @@ def _window_setups(
             stop_loss=stop,
             entry_price=entry,
         )
+        # meets_core = clears every gate EXCEPT target probability, so the report
+        # can sweep the min_target_probability threshold without re-replaying.
+        core_config = {**activation, "min_target_probability": 0.0}
+        meets_core = setup_qualifies(setup_ns, core_config)
+        best_prob = best_target_probability(setup_ns)
+        min_tp = float(activation.get("min_target_probability", 0.0))
         out.append({
             "direction": direction,
             "entry": entry,
@@ -168,10 +174,11 @@ def _window_setups(
             "rr": rr,
             "confidence": confidences[direction],
             "primary_prob": float(primary["probability"]),
-            "best_prob": best_target_probability(setup_ns),
+            "best_prob": best_prob,
+            "meets_core": meets_core,
             "action": action,
             "risk_level": risk_level,
-            "qualified": setup_qualifies(setup_ns, activation),
+            "qualified": meets_core and best_prob >= min_tp,
         })
     return out
 
@@ -208,6 +215,8 @@ def _replay_ticker(symbol: str, records: list, config: dict, activation: dict) -
                 "rr": s["rr"],
                 "confidence": s["confidence"],
                 "primary_prob": s["primary_prob"],
+                "best_prob": s["best_prob"],
+                "meets_core": s["meets_core"],
                 "qualified": s["qualified"],
                 "outcome": outcome,
                 "target_hit": target_hit,
@@ -279,6 +288,15 @@ async def run_backtest(
     longs = [c for c in qualified if c["direction"] == "long"]
     shorts = [c for c in qualified if c["direction"] == "short"]
 
+    # Threshold sweep: re-apply the gate at several min_target_probability values
+    # (holding the other conditions fixed) so the trade-off between how many
+    # setups qualify and their expectancy is visible without re-replaying.
+    current_min_tp = float(activation.get("min_target_probability", 60.0))
+    sweep = []
+    for threshold in (60, 55, 50, 45, 40, 35, 30):
+        cands = [c for c in candidates if c["meets_core"] and c["best_prob"] >= threshold]
+        sweep.append({"min_target_probability": threshold, **_bucket_stats(cands)})
+
     return {
         "generated_at": datetime.now(timezone.utc).isoformat(),
         "tickers": total,
@@ -292,6 +310,8 @@ async def run_backtest(
             "long": _bucket_stats(longs),
             "short": _bucket_stats(shorts),
         },
+        "min_target_probability": current_min_tp,
+        "sweep": sweep,
         "calibration": _calibration(candidates),
         "note": (
             "Sentiment & fundamentals held neutral (no point-in-time history). "

frontend/src/components/signals/BacktestPanel.tsx

@@ -158,6 +158,53 @@ export function BacktestPanel() {
               </table>
             </div>
 
+            {report.sweep && report.sweep.length > 0 && (
+              <div>
+                <p className="mb-2 text-xs font-medium uppercase tracking-widest text-gray-500">
+                  Min target-probability sweep
+                </p>
+                <p className="mb-2 text-[11px] text-gray-500">
+                  How many setups qualify — and how they perform — at each gate threshold (other
+                  gate conditions held fixed). Lower = more trades, watch that expectancy holds.
+                  Your current setting is highlighted; set it in Admin → Settings → Activation.
+                </p>
+                <div className="glass overflow-x-auto">
+                  <table className="w-full text-sm">
+                    <thead>
+                      <tr className="border-b border-white/[0.06] text-left text-xs uppercase tracking-wider text-gray-500">
+                        <th className="px-4 py-2.5">Min Target Prob</th>
+                        <th className="px-4 py-2.5 text-right">Qualified</th>
+                        <th className="px-4 py-2.5 text-right">Wins</th>
+                        <th className="px-4 py-2.5 text-right">Losses</th>
+                        <th className="px-4 py-2.5 text-right">Hit Rate</th>
+                        <th className="px-4 py-2.5 text-right">Avg R</th>
+                        <th className="px-4 py-2.5 text-right">Total R</th>
+                      </tr>
+                    </thead>
+                    <tbody>
+                      {report.sweep.map((row) => {
+                        const current = Math.abs(row.min_target_probability - report.min_target_probability) < 0.5;
+                        return (
+                          <tr key={row.min_target_probability} className={`border-b border-white/[0.04] ${current ? 'bg-blue-400/10' : ''}`}>
+                            <td className="num px-4 py-2.5 text-gray-200">
+                              {current && <span className="mr-1 text-blue-300">★</span>}
+                              {row.min_target_probability}%
+                            </td>
+                            <td className="num px-4 py-2.5 text-right text-gray-200">{row.total}</td>
+                            <td className="num px-4 py-2.5 text-right text-emerald-400">{row.wins}</td>
+                            <td className="num px-4 py-2.5 text-right text-red-400">{row.losses}</td>
+                            <td className="num px-4 py-2.5 text-right text-gray-200">{fmtPct(row.hit_rate)}</td>
+                            <td className={`num px-4 py-2.5 text-right font-semibold ${rColor(row.avg_r)}`}>{fmtR(row.avg_r)}</td>
+                            <td className={`num px-4 py-2.5 text-right ${rColor(row.total_r)}`}>{fmtR(row.total_r)}</td>
+                          </tr>
+                        );
+                      })}
+                    </tbody>
+                  </table>
+                </div>
+              </div>
+            )}
+
             <div>
               <p className="mb-2 text-xs font-medium uppercase tracking-widest text-gray-500">
                 Probability calibration

frontend/src/lib/types.ts

@@ -196,6 +196,10 @@ export interface BacktestCalibrationRow {
   realized_hit_rate: number;
 }
 
+export interface BacktestSweepRow extends BacktestBucket {
+  min_target_probability: number;
+}
+
 export interface BacktestReport {
   generated_at: string;
   tickers: number;
@@ -205,6 +209,8 @@ export interface BacktestReport {
   overall_qualified: BacktestBucket;
   overall_all: BacktestBucket;
   by_direction: Record<string, BacktestBucket>;
+  min_target_probability: number;
+  sweep: BacktestSweepRow[];
   calibration: BacktestCalibrationRow[];
   note: string;
 }

tests/unit/test_backtest_service.py

@@ -108,10 +108,15 @@ async def test_run_backtest_smoke(session):
     # well-formed report
     assert report["tickers"] == 1
     assert isinstance(report["candidates"], int)
-    for key in ("overall_qualified", "overall_all", "by_direction", "calibration"):
+    for key in ("overall_qualified", "overall_all", "by_direction", "calibration", "sweep"):
         assert key in report
     # the oscillating series should yield at least some resolved setups
     assert report["candidates"] >= 1
+
+    # sweep: lowering the threshold can only add qualifiers, never remove them
+    sweep = sorted(report["sweep"], key=lambda r: r["min_target_probability"], reverse=True)
+    counts = [r["total"] for r in sweep]
+    assert counts == sorted(counts)  # ascending as threshold descends
     # every calibration row is internally consistent
     for row in report["calibration"]:
         assert 0 <= row["realized_hit_rate"] <= 100