diff --git a/deploy/profile-scripts/b_td1_v3_gen.py b/deploy/profile-scripts/b_td1_v3_gen.py
new file mode 100644
index 00000000..bde21577
--- /dev/null
+++ b/deploy/profile-scripts/b_td1_v3_gen.py
@@ -0,0 +1,112 @@
+# -*- coding: utf-8 -*-
+"""b_td1_v3 生成器（基于 DB 原版信号 + merge 查 score + 每日top-N截断）"""
+import sys, json
+sys.path.insert(0, "/home/hmo/MoFin")
+import pandas as pd
+import numpy as np
+import sqlite3
+
+TOP_N = 5
+SRC_VERSION = "b_td1"
+
+
+def score_row(b, r, s, r5):
+    sc = 0
+    if b is not None and not pd.isna(b): sc += 40 if b < -30 else 32 if b < -20 else 20 if b < -10 else 8
+    if r is not None and not pd.isna(r): sc += 30 if r < 30 else 24 if r < 40 else 14 if r < 50 else 6
+    if s is not None and not pd.isna(s): sc += 20 if s < -20 else 14 if s < -10 else 8 if s < 0 else 3
+    if r5 is not None and not pd.isna(r5): sc += 10 if r5 < -25 else 7 if r5 < -15 else 4 if r5 < -8 else 1
+    return sc
+
+
+def gen_trades(start, end, top_n=TOP_N):
+    conn = sqlite3.connect("/home/hmo/MoFin/data/mofin.db", timeout=30)
+    r = conn.execute("SELECT results_json FROM strategy_research WHERE version=? AND period_tag='2y'",
+                     (SRC_VERSION,)).fetchone()
+    conn.close()
+    if not r:
+        return []
+    d = json.loads(r[0])
+    trades = [t for t in d["trades"] if start <= t["entry_date"] <= end]
+    if not trades:
+        return []
+
+    # panel 特征用 merge（快）
+    panel = pd.read_pickle("/tmp/panel_12d.pkl")
+    pcols = ["bias60", "rsi", "sec_ret20", "ret5"]
+    pf = panel[["code", "date"] + pcols].copy()
+    pf["code"] = pf["code"].astype(str).str.zfill(6)
+    pf["date"] = pf["date"].astype(str)
+
+    tdf = pd.DataFrame(trades)
+    tdf["code"] = tdf["code"].astype(str).str.zfill(6)
+    tdf["date"] = tdf["entry_date"].astype(str)
+    merged = tdf[["code", "date"]].merge(pf, on=["code", "date"], how="left")
+    merged["score"] = merged.apply(lambda r: score_row(r["bias60"], r["rsi"], r["sec_ret20"], r["ret5"]), axis=1)
+    score_map = dict(zip(zip(merged["code"], merged["date"]), merged["score"]))
+    for t in trades:
+        t["score"] = score_map.get((str(t["code"]).zfill(6), t["entry_date"]), 50)
+        t.setdefault("boost", 1.0)
+
+    # 每日 top-N
+    from collections import defaultdict
+    by_day = defaultdict(list)
+    for t in trades:
+        by_day[t["entry_date"]].append(t)
+    chosen = []
+    for dt in sorted(by_day):
+        chosen.extend(sorted(by_day[dt], key=lambda x: -x["score"])[:top_n])
+    return chosen
+
+
+def build_results(trades):
+    import copy, random
+    from strategy_lab import portfolio_sim
+    n = len(trades)
+    wins = [t for t in trades if t["profit_pct"] > 0]
+    wr = len(wins) / n * 100 if n else 0
+    avg = sum(t["profit_pct"] for t in trades) / n if n else 0
+    sim = portfolio_sim(trades, 1000000, max_positions=10) or {}
+    rets = []
+    for seed in range(5):
+        t2 = copy.deepcopy(trades); rng = random.Random(seed); rng.shuffle(t2)
+        rets.append((portfolio_sim(t2, 1000000, max_positions=10) or {}).get("total_return_pct") or 0)
+    return {"summary": {"total_trades": n, "win_rate": round(wr, 1), "avg_profit_pct": round(avg, 2)},
+            "portfolio": {"positions_taken": sim.get("positions_taken"), "positions_skipped": sim.get("positions_skipped"),
+                          "total_return_pct": sim.get("total_return_pct"), "cagr_pct": sim.get("cagr_pct"),
+                          "portfolio_max_dd_pct": sim.get("portfolio_max_dd_pct"), "capital_final": sim.get("capital_final")},
+            "robustness": {"shuffle_total_return": rets, "spread_pp": round(max(rets) - min(rets), 1)},
+            "trades": trades}
+
+
+if __name__ == "__main__":
+    from datetime import datetime
+    DB = "/home/hmo/MoFin/data/mofin.db"
+    VERSION = "b_td1_v3"
+    windows = {"1y": ("2025-07-01", "2026-07-01"), "2y": ("2024-07-01", "2026-07-01"),
+               "5y": ("2021-07-01", "2026-07-01"), "10y": ("2016-01-01", "2026-07-01")}
+    conn = sqlite3.connect(DB, timeout=30)
+    now = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
+    for pt, (s, e) in windows.items():
+        trades = gen_trades(s, e)
+        if not trades:
+            print(f"[{pt}] 无信号")
+            continue
+        results = build_results(trades)
+        exist = conn.execute("SELECT id FROM strategy_research WHERE version=? AND period_tag=?", (VERSION, pt)).fetchone()
+        if exist:
+            conn.execute("UPDATE strategy_research SET results_json=? WHERE version=? AND period_tag=?",
+                         (json.dumps(results, ensure_ascii=False), VERSION, pt))
+        else:
+            conn.execute("""INSERT INTO strategy_research (version, name, summary, hypothesis, parent, config_json, results_json, period, created_at, market, period_tag, deprecated)
+                VALUES (?,?,?,?,?,?,?,?,?,?,?,?)""",
+                (VERSION, "B组超跌·原池优选", "b_td1原池+score每日top5截断",
+                 "老莫指正: v2超跌强化错误砍成交。改原池+score每日top5 → 信号-89%/成交-9%/比1.8",
+                 "B组", json.dumps({"top_n": TOP_N, "src": SRC_VERSION}, ensure_ascii=False),
+                 json.dumps(results, ensure_ascii=False), None, now, "a", pt, None))
+        s2 = results["summary"]; p = results["portfolio"] or {}
+        _pt2 = p.get("positions_taken") or 0
+        print(f"[{pt}] 信号{s2['total_trades']} 成交{_pt2} 比{s2['total_trades']/max(_pt2,1):.1f} "
+              f"胜率{s2['win_rate']}% 年化{p.get('cagr_pct') or 0}% 回撤{p.get('portfolio_max_dd_pct') or 0}% 洗牌差{results['robustness']['spread_pp']}pp")
+    conn.commit(); conn.close()
+    print(f"\n写入完成 (version={VERSION})")