1

AGENTS.md

@@ -107,6 +107,7 @@ AlphaX 是一个以 `Python + FastAPI + PostgreSQL + Docker + 静态 HTML` 组
 - `signal_performance` 是复盘后动态权重来源；`review_engine.py` 更新信号绩效后，`config_loader.get_signal_weights()` 会让下一轮筛选/确认读取生效权重。
 - 当前确认层已把核心技术因子、资金面因子、板块因子、舆情因子和买点风险因子接入 `FactorScorer`，并在 `market_context.factor_score_breakdown` / `entry_plan.factor_score_breakdown` 中保留因子明细。
 - `FactorScorer` 已加入因子组去相关，同一类 `momentum` / `structure` / `entry_quality` / `onchain_flow` / `narrative` 信号会受 group cap 限制，避免同一根行情被重复加分。
+- 小样本复盘不能直接杀死核心因子。`signal_performance` 的动态权重至少要满足 `review.min_samples_for_weight` 与 `review.signal_deprecation.min_samples` 后才覆盖确认层基线；未达样本门槛时只用于观察，不应用 0 权重把 15min 启动、日线突破回踩等因子压没。
 - 扣分因子应传负数，例如 `FactorScorer.delta("false_breakout", -5, ...)`，不要再外部 `score -= delta`，否则 `factor_score_breakdown` 会把风险误记成正向贡献。
 - 确认层会输出 `score_components`：`opportunity_score` 表示机会质量，`entry_score` 表示买点质量，`risk_score` 表示扣分风险；后续策略不要再只看单一 `rec_score`。
 - `market_context.decision_log` / `entry_plan.decision_log` 是结构化决策解释；paper trading 开仓事件也会记录当时 `market_regime`、`global_risk` 和 `score_components`。
@@ -163,6 +164,7 @@ AlphaX 是一个以 `Python + FastAPI + PostgreSQL + Docker + 静态 HTML` 组
   - 市场环境识别中心，第一版基于市场快照、BTC/ETH 涨跌、山寨涨跌广度、强势/大跌数量和 funding 热度识别 `risk_off`、`btc_main_uptrend`、`altcoin_rotation`、`sideways_chop`、`meme_frenzy`、`unknown`。
 - `app/core/global_risk.py`
   - paper trading 全局风控门禁。单币机会进入开仓或挂单成交前，需要先检查市场环境和账户风险；critical 禁止新开仓，high 只允许高质量机会。
+  - 确认层也会应用同一市场风控语义：`risk_level=critical` 且 `position_multiplier=0` 时，强势发现仍可记录为观察，但不能输出 `buy_now` 或新挂单动作；已有活跃可交易推荐会被降级为观察并写入 `market_risk_gate`。
 
 ## 5. 数据与状态中心
 

app/config/config_loader.py

@@ -267,7 +267,10 @@ def get_signal_weights():
     - 旧脚本仍可能用历史 key（如 "1H放量(量价背离)"）直接查 weights[...] 
     因此返回值同时暴露 canonical key + alias key，避免旧调用方 KeyError。
     """
-    rules = load_rules()
+    # Signal weights need a stable baseline. Runtime strategy_override may
+    # contain small-sample governance writes; those are only trusted through
+    # signal_performance after the sample-size gate below.
+    rules = _load_yaml_baseline()
     yaml_weights = copy.deepcopy(rules.get("signal_weights", {}))
 
     canonical = {}
@@ -277,9 +280,15 @@ def get_signal_weights():
     try:
         from app.db.altcoin_db import get_signal_weights as db_get_weights
         db_weights = db_get_weights()
+        review_params = get_review_params()
+        deprecation_params = review_params.get("signal_deprecation") or {}
+        min_samples = max(
+            int(review_params.get("min_samples_for_weight", 12) or 12),
+            int(deprecation_params.get("min_samples", 12) or 12),
+        )
         for sig, data in db_weights.items():
             norm_sig = normalize_signal_name(sig)
-            if data.get("total_count", 0) >= 3:
+            if data.get("total_count", 0) >= min_samples:
                 canonical[norm_sig] = data["weight"]
     except Exception:
         pass

app/db/altcoin_db.py

@@ -70,6 +70,11 @@ def expire_old_recommendations(*args, **kwargs):
     return _recommendation_commands.expire_old_recommendations(*args, **kwargs)
 
 
+def downgrade_active_entries_for_market_risk(*args, **kwargs):
+    _sync_command_compat_hooks()
+    return _recommendation_commands.downgrade_active_entries_for_market_risk(*args, **kwargs)
+
+
 def apply_recommendation_state_transition(*args, **kwargs):
     _sync_command_compat_hooks()
     return _recommendation_commands.apply_recommendation_state_transition(*args, **kwargs)

app/db/paper_trading.py

@@ -1390,6 +1390,138 @@ def sync_recommendation(rec: dict, current_price: float, event_time: str = "") -
             pass
 
 
+def sync_pending_paper_orders(limit: int = 100, event_time: str = "", config: dict | None = None) -> dict:
+    """Reconcile pending limit orders against the latest shared price cache.
+
+    The strategy runner can miss an existing order if the recommendation is
+    later derived back into observe status. Pending orders are executable
+    state, so they need their own reconciliation pass based on the same
+    latest_price_cache used by the Web page.
+    """
+    if not paper_trading_enabled():
+        return {"enabled": False, "processed_count": 0, "results": []}
+    limit = max(1, min(_safe_int(limit, 100), 500))
+    event_time = event_time or _now()
+    cfg = _paper_cfg(config)
+    conn = get_conn()
+    results = []
+    try:
+        rows = conn.execute(
+            """
+            SELECT
+                po.*,
+                r.status AS rec_status,
+                r.rec_state,
+                r.rec_score,
+                r.action_status,
+                r.execution_status,
+                r.lifecycle_state,
+                r.display_bucket,
+                r.entry_price,
+                r.current_price AS rec_current_price,
+                r.stop_loss AS rec_stop_loss,
+                r.tp1 AS rec_tp1,
+                r.tp2 AS rec_tp2,
+                r.strategy_version AS rec_strategy_version,
+                r.entry_plan_json,
+                r.market_context_json,
+                r.derivatives_context_json,
+                r.sector_context_json,
+                lpc.price AS latest_price,
+                lpc.updated_at AS latest_price_updated_at
+            FROM paper_orders po
+            LEFT JOIN recommendation r ON r.id = po.recommendation_id
+            LEFT JOIN latest_price_cache lpc ON lpc.symbol = po.symbol
+            WHERE po.status='pending'
+            ORDER BY po.created_at ASC, po.id ASC
+            LIMIT %s
+            """,
+            (limit,),
+        ).fetchall()
+        for row in rows:
+            item = dict(row)
+            current_price = _safe_float(item.get("latest_price"))
+            order = {k: item.get(k) for k in [
+                "id",
+                "recommendation_id",
+                "symbol",
+                "side",
+                "order_type",
+                "status",
+                "source_status",
+                "source_action",
+                "target_price",
+                "current_price_at_create",
+                "fill_price",
+                "notional_usdt",
+                "stop_loss",
+                "tp1",
+                "tp2",
+                "strategy_version",
+                "entry_plan_snapshot_json",
+                "created_at",
+                "updated_at",
+                "expires_at",
+                "filled_at",
+                "canceled_at",
+                "cancel_reason",
+            ]}
+            rec = {
+                "id": item.get("recommendation_id"),
+                "symbol": item.get("symbol"),
+                "status": item.get("rec_status") or "active",
+                "rec_state": item.get("rec_state"),
+                "rec_score": item.get("rec_score"),
+                "action_status": item.get("action_status") or item.get("source_action"),
+                "execution_status": item.get("execution_status") or item.get("source_status"),
+                "lifecycle_state": item.get("lifecycle_state"),
+                "display_bucket": item.get("display_bucket"),
+                "entry_price": item.get("entry_price") or item.get("target_price"),
+                "current_price": item.get("rec_current_price") or current_price,
+                "stop_loss": item.get("rec_stop_loss") or item.get("stop_loss"),
+                "tp1": item.get("rec_tp1") or item.get("tp1"),
+                "tp2": item.get("rec_tp2") or item.get("tp2"),
+                "strategy_version": item.get("rec_strategy_version") or item.get("strategy_version"),
+                "entry_plan_json": item.get("entry_plan_json") or item.get("entry_plan_snapshot_json"),
+                "market_context_json": item.get("market_context_json"),
+                "derivatives_context_json": item.get("derivatives_context_json"),
+                "sector_context_json": item.get("sector_context_json"),
+            }
+            if current_price <= 0:
+                result = {
+                    "skipped": True,
+                    "reason": "missing_latest_price",
+                    "paper_order_id": order.get("id"),
+                    "symbol": order.get("symbol"),
+                    "target_price": order.get("target_price"),
+                }
+            else:
+                result = _sync_wait_pullback_order(conn, rec, current_price, event_time, cfg)
+                result.update({
+                    "symbol": order.get("symbol"),
+                    "latest_price": current_price,
+                    "latest_price_updated_at": item.get("latest_price_updated_at") or "",
+                })
+            results.append(result)
+        conn.commit()
+    except Exception:
+        conn.rollback()
+        raise
+    finally:
+        try:
+            conn.close()
+        except Exception:
+            pass
+    return {
+        "enabled": True,
+        "processed_count": len(results),
+        "filled_count": sum(1 for r in results if r.get("paper_order", {}).get("filled")),
+        "canceled_count": sum(1 for r in results if str(r.get("reason") or "").startswith("paper_order_") and "canceled" in str(r.get("reason") or "")),
+        "results": results,
+        "run_time": event_time,
+    }
+
+
 def get_paper_trading_summary(days: int = 30) -> dict:
     days = max(1, min(_safe_int(days, 30), 365))
     cutoff = (datetime.now() - timedelta(days=days)).isoformat()

app/db/recommendation_commands.py

@@ -244,6 +244,66 @@ def expire_old_recommendations(hours=48):
     conn.close()
 
 
+def downgrade_active_entries_for_market_risk(reason: str, event_time: str | None = None) -> dict:
+    """Downgrade active executable recommendations when global market risk blocks entries."""
+    event_time = event_time or datetime.now().isoformat()
+    reason = str(reason or "全市场风险过高，暂停新开仓与新挂单").strip()
+    conn = get_conn()
+    rows = conn.execute(
+        """
+        SELECT id, entry_plan_json, action_status
+        FROM recommendation
+        WHERE status='active'
+          AND (
+            COALESCE(action_status,'') IN ('可即刻买入','等回踩')
+            OR COALESCE(execution_status,'') IN ('buy_now','wait_pullback')
+          )
+        ORDER BY id DESC
+        """
+    ).fetchall()
+    updated = 0
+    execution_status, display_bucket, lifecycle_state, entry_triggered, state_reason = derive_minimal_state_fields(
+        "active", "观察", {}
+    )
+    for row in rows:
+        entry_plan = normalize_json_object(row["entry_plan_json"])
+        previous_action = str(row["action_status"] or "").strip()
+        entry_plan["market_risk_gate"] = {
+            "blocked_action": previous_action,
+            "final_action": "观察",
+            "risk_level": "critical",
+            "reasons": [reason],
+            "updated_at": event_time,
+        }
+        entry_plan["entry_action"] = "观察"
+        conn.execute(
+            """
+            UPDATE recommendation
+            SET action_status='观察',
+                execution_status=%s,
+                display_bucket=%s,
+                lifecycle_state=%s,
+                entry_triggered=%s,
+                state_reason=%s,
+                entry_plan_json=%s
+            WHERE id=%s
+            """,
+            (
+                execution_status,
+                display_bucket,
+                lifecycle_state,
+                entry_triggered,
+                state_reason,
+                json.dumps(entry_plan, ensure_ascii=False),
+                row["id"],
+            ),
+        )
+        updated += 1
+    conn.commit()
+    conn.close()
+    return {"updated_count": updated, "reason": reason, "updated_at": event_time}
+
+
 def apply_recommendation_state_transition(rec_id, requested_action, current_price, event_time=None, signals=None):
     """The single DB entry for turning price events into recommendation action state."""
     event_time = event_time or datetime.now().strftime("%Y-%m-%dT%H:%M:%S")

app/services/altcoin_confirm.py

@@ -29,7 +29,7 @@ from app.core.sector_map import get_burst_threshold, is_meme_coin, get_sector_fo
 from app.db.altcoin_db import (
     init_db, expire_old_states, expire_old_recommendations,
     get_candidates_for_confirm, update_state, get_conn, create_recommendation, log_screening,
-    log_cron_run, update_latest_price_cache,
+    log_cron_run, update_latest_price_cache, downgrade_active_entries_for_market_risk,
 )
 from app.config.config_loader import (
     get_strategy_direction,
@@ -524,6 +524,32 @@ def _decision_log(module: str, decision: str, *, score: float = 0.0, reasons=Non
     }
 
 
+def _apply_market_risk_entry_gate(entry_plan: dict, signals: list, market_regime: dict) -> tuple[dict, str]:
+    """Keep high-risk market discoveries visible, but block executable entries."""
+    plan = dict(entry_plan or {})
+    regime = market_regime if isinstance(market_regime, dict) else {}
+    risk_level = str(regime.get("risk_level") or "").strip().lower()
+    position_multiplier = float(regime.get("position_multiplier") or 0)
+    current_action = str(plan.get("entry_action") or "").strip()
+    if risk_level != "critical" or position_multiplier > 0:
+        return plan, ""
+    if current_action not in {"可即刻买入", "即刻买入", "等回踩"}:
+        return plan, ""
+
+    reason = "全市场处于 critical 风险，暂停新开仓与新挂单，保留为观察机会"
+    plan["market_risk_gate"] = {
+        "blocked_action": current_action,
+        "final_action": "观察",
+        "risk_level": risk_level,
+        "position_multiplier": position_multiplier,
+        "reasons": [reason],
+    }
+    plan["entry_action"] = "观察"
+    if not any("市场风控闸门" in str(sig) for sig in signals):
+        signals.append(f"⚠️ 市场风控闸门: {reason}")
+    return plan, reason
+
+
 # ==================== 确认逻辑 ====================
 
 def detect_volume_price_fly_1h(df_1h):
@@ -1437,6 +1463,15 @@ def confirm_burst(symbol, cand):
     sector_context = compute_sector_context(symbol, cand_detail)
     regime_context = _current_market_regime_context()
     market_regime = regime_context.get("market_regime") or {}
+    if entry_plan:
+        entry_plan, market_risk_gate_reason = _apply_market_risk_entry_gate(entry_plan, signals, market_regime)
+        if market_risk_gate_reason:
+            score += factor_scorer.delta(
+                "entry_quality_gate",
+                -2,
+                evidence="全局市场风险闸门降为观察",
+                value=market_risk_gate_reason,
+            )
     factor_score_breakdown = factor_scorer.summary()
     opportunity_score = round(float(factor_score_breakdown.get("opportunity_score") or 0), 3)
     entry_score = round(float(factor_score_breakdown.get("entry_score") or 0), 3)
@@ -1498,7 +1533,7 @@ def confirm_burst(symbol, cand):
         "pa_15min": pa_15min_result,
         "pa_1d": pa_1d,
         "m30_aligned": m30_aligned,
-        "entry_action": entry_action,
+        "entry_action": (entry_plan or {}).get("entry_action") or entry_action,
         "market_context": market_context,
         "derivatives_context": derivatives_context,
         "sector_context": sector_context,
@@ -1584,6 +1619,15 @@ def main(compact: bool = False):
     try:
         init_db()
         expire_old_states()
+        regime_context = _current_market_regime_context()
+        market_regime = regime_context.get("market_regime") or {}
+        if str(market_regime.get("risk_level") or "").strip().lower() == "critical" and float(market_regime.get("position_multiplier") or 0) <= 0:
+            downgrade_result = downgrade_active_entries_for_market_risk(
+                "全市场处于 critical 风险，暂停新开仓与新挂单，保留为观察机会",
+                event_time=datetime.now().isoformat(),
+            )
+        else:
+            downgrade_result = {"updated_count": 0}
 
         candidates = get_candidates_for_confirm()
 
@@ -1591,6 +1635,7 @@ def main(compact: bool = False):
             output = {
                 "status": "no_candidates",
                 "message": "无需要确认的候选（需加速状态+评分≥6）",
+                "market_risk_downgraded_count": downgrade_result.get("updated_count", 0),
                 "check_time": datetime.now().isoformat(),
             }
             _emit_output(output, compact=compact)
@@ -1734,6 +1779,7 @@ def main(compact: bool = False):
             "status": "confirmed" if confirmed else "unconfirmed",
             "confirmed_count": len(confirmed),
             "unconfirmed_count": len(unconfirmed),
+            "market_risk_downgraded_count": downgrade_result.get("updated_count", 0),
             "confirmed": confirmed,
             "unconfirmed": unconfirmed,
             "check_time": datetime.now().isoformat(),

app/services/paper_trader.py

@@ -8,7 +8,7 @@ from datetime import datetime
 import ccxt
 
 from app.db.altcoin_db import init_db, log_cron_run, update_latest_price_cache
-from app.db.paper_trading import get_paper_trading_summary, sync_recommendation
+from app.db.paper_trading import get_paper_trading_summary, sync_pending_paper_orders, sync_recommendation
 from app.db.recommendation_queries import get_active_recommendations_deduped
 from app.services.live_trading_sync import sync_paper_trade_to_live
 
@@ -35,12 +35,19 @@ def run_once(limit: int = 100) -> dict:
             results.append(result)
         except Exception as exc:
             failed.append({"symbol": symbol, "error": str(exc)})
+    pending_result = sync_pending_paper_orders(limit=limit, event_time=datetime.now().isoformat())
+    for item in pending_result.get("results", []):
+        if item.get("trade_id") and (item.get("opened") or item.get("paper_order", {}).get("filled")):
+            item["live_sync"] = sync_paper_trade_to_live(int(item["trade_id"]), execute=True)
     output = {
         "status": "completed",
         "processed_count": len(results),
+        "pending_processed_count": pending_result.get("processed_count", 0),
+        "pending_filled_count": pending_result.get("filled_count", 0),
         "failed_count": len(failed),
         "failed": failed,
         "results": results,
+        "pending_results": pending_result.get("results", []),
         "summary": get_paper_trading_summary(days=30),
         "run_time": datetime.now().isoformat(),
     }
@@ -77,6 +84,8 @@ def main(limit: int = 100):
         duration_ms=int((finished_at - started_at).total_seconds() * 1000),
         summary={
             "processed_count": output.get("processed_count", 0),
+            "pending_processed_count": output.get("pending_processed_count", 0),
+            "pending_filled_count": output.get("pending_filled_count", 0),
             "failed_count": output.get("failed_count", 0),
             "open_count": output.get("summary", {}).get("open_count", 0),
             "closed_count": output.get("summary", {}).get("closed_count", 0),

app/services/review_engine.py

@@ -87,10 +87,10 @@ def _get_thresholds():
         "hit_threshold_pct": params.get("hit_threshold_pct", 5.0),
         "fail_threshold_pct": params.get("fail_threshold_pct", -3.0),
         "missed_explosion_pct": params.get("missed_explosion_pct", 20.0),
-        "min_samples_for_weight": params.get("min_samples_for_weight", 3),
+        "min_samples_for_weight": params.get("min_samples_for_weight", 12),
         "weight_floor": params.get("weight_floor", 0.0),
         "hit_rate_kill_threshold": params.get("hit_rate_kill_threshold", 0.10),
-        "kill_min_samples": params.get("kill_min_samples", 5),
+        "kill_min_samples": params.get("kill_min_samples", 20),
         "category_base_weights": params.get("category_base_weights", {"前瞻": 2.0, "PA": 1.5, "滞后": 0.5}),
     }
 
@@ -605,8 +605,8 @@ def _apply_daily_factor_weight_governance():
     """
     thresholds = _get_thresholds()
     weights = get_signal_weights()
-    min_samples = max(3, int(thresholds.get("min_samples_for_weight", 3) or 3))
+    min_samples = max(12, int(thresholds.get("min_samples_for_weight", 12) or 12))
-    kill_min_samples = max(min_samples, int(thresholds.get("kill_min_samples", 5) or 5))
+    kill_min_samples = max(min_samples, int(thresholds.get("kill_min_samples", 20) or 20))
     kill_hit_rate = float(thresholds.get("hit_rate_kill_threshold", 0.10) or 0.10) * 100
     warn_hit_rate = float((thresholds.get("signal_deprecation") or {}).get("hit_rate_warn_threshold", 0.20) or 0.20) * 100
     category_base = thresholds.get("category_base_weights") or {"前瞻": 2.0, "PA": 1.5, "滞后": 0.5}

rules.yaml

@@ -271,21 +271,21 @@ review:
   hit_threshold_pct: 5.0
   fail_threshold_pct: -3.0
   missed_explosion_pct: 20.0
-  min_samples_for_weight: 3
+  min_samples_for_weight: 12
   weight_floor: 0.0
   hit_rate_kill_threshold: 0.1
-  kill_min_samples: 5
+  kill_min_samples: 20
   category_base_weights:
     前瞻: 2.0
     PA: 1.5
     滞后: 0.5
   signal_deprecation:
     enabled: true
-    min_samples: 10
+    min_samples: 20
     hit_rate_warn_threshold: 0.2
     hit_rate_deprecate_threshold: 0.1
     min_tracking_days: 3
-    note: 信号淘汰机制。累计≥10样本+跟踪≥3天后，命中率<10%自动淘汰(权重→0)；<20%警告并自动降权。
+    note: 信号淘汰机制。累计≥20样本+跟踪≥3天后，命中率<10%自动淘汰(权重→0)；<20%警告并自动降权；小样本只观察，不直接影响确认层权重。
     deprecate_action: set_weight_zero
     warn_action: halve_weight
 reverse_analysis:

tests/test_confirm_market_risk_gate.py

@@ -0,0 +1,28 @@
+from app.services.altcoin_confirm import _apply_market_risk_entry_gate
+
+
+def test_market_risk_gate_blocks_executable_action_when_critical():
+    signals = ["15min即刻入场信号"]
+    plan, reason = _apply_market_risk_entry_gate(
+        {"entry_action": "可即刻买入", "entry_price": 1.0},
+        signals,
+        {"risk_level": "critical", "position_multiplier": 0.0},
+    )
+
+    assert plan["entry_action"] == "观察"
+    assert plan["market_risk_gate"]["blocked_action"] == "可即刻买入"
+    assert "暂停新开仓" in reason
+    assert any("市场风控闸门" in sig for sig in signals)
+
+
+def test_market_risk_gate_keeps_observation_action_visible():
+    signals = []
+    plan, reason = _apply_market_risk_entry_gate(
+        {"entry_action": "观察", "entry_price": 1.0},
+        signals,
+        {"risk_level": "critical", "position_multiplier": 0.0},
+    )
+
+    assert plan["entry_action"] == "观察"
+    assert "market_risk_gate" not in plan
+    assert reason == ""

tests/test_factor_scoring.py

@@ -45,10 +45,41 @@ def test_factor_scorer_records_negative_risk_as_risk_score():
 
 
 def test_signal_weight_alias_keeps_legacy_chinese_keys_available(monkeypatch):
-    monkeypatch.setattr("app.config.config_loader.load_rules", lambda: {"signal_weights": {"量价齐飞": 5}})
+    monkeypatch.setattr("app.config.config_loader._load_yaml_baseline", lambda: {"signal_weights": {"量价齐飞": 5}})
     monkeypatch.setattr("app.db.altcoin_db.get_signal_weights", lambda: {})
 
     weights = get_signal_weights()
 
     assert weights["vp_fly_1h_current"] == 5
     assert weights["量价齐飞"] == 5
+
+
+def test_signal_weight_ignores_thin_dynamic_samples(monkeypatch):
+    monkeypatch.setattr(
+        "app.config.config_loader._load_yaml_baseline",
+        lambda: {"signal_weights": {"量价齐飞": 5}},
+    )
+    monkeypatch.setattr(
+        "app.config.config_loader.load_rules",
+        lambda: {
+            "review": {
+                "min_samples_for_weight": 12,
+                "signal_deprecation": {"min_samples": 20},
+            },
+        },
+    )
+    monkeypatch.setattr(
+        "app.db.altcoin_db.get_signal_weights",
+        lambda: {
+            "vp_fly_1h_current": {
+                "weight": 0,
+                "total_count": 3,
+                "hit_rate": 0,
+                "avg_pnl": -1,
+            }
+        },
+    )
+
+    weights = get_signal_weights()
+
+    assert weights["vp_fly_1h_current"] == 5

tests/test_market_risk_recommendation_downgrade.py

@@ -0,0 +1,39 @@
+import json
+
+from app.db.recommendation_commands import downgrade_active_entries_for_market_risk
+
+
+def test_downgrade_active_entries_for_market_risk_updates_executable_recs(pg_conn):
+    pg_conn.execute(
+        """
+        INSERT INTO recommendation (
+            symbol, rec_time, rec_state, rec_score, entry_price, status,
+            action_status, execution_status, display_bucket, lifecycle_state,
+            entry_plan_json
+        )
+        VALUES
+            ('BUY/USDT', '2026-05-26T10:00:00', '爆发', 30, 1.0, 'active',
+             '可即刻买入', 'buy_now', 'realtime', 'buyable', '{"entry_action":"可即刻买入"}'),
+            ('WAIT/USDT', '2026-05-26T10:01:00', '爆发', 20, 1.0, 'active',
+             '等回踩', 'wait_pullback', 'watch_pool', 'waiting_entry', '{"entry_action":"等回踩"}'),
+            ('OBS/USDT', '2026-05-26T10:02:00', '观察', 10, 1.0, 'active',
+             '观察', 'observe', 'watch_pool', 'watching', '{"entry_action":"观察"}')
+        """
+    )
+    pg_conn.commit()
+
+    result = downgrade_active_entries_for_market_risk("critical test", event_time="2026-05-26T11:00:00")
+
+    assert result["updated_count"] == 2
+    rows = pg_conn.execute(
+        "SELECT symbol, action_status, execution_status, display_bucket, entry_plan_json FROM recommendation ORDER BY symbol"
+    ).fetchall()
+    by_symbol = {row["symbol"]: dict(row) for row in rows}
+    assert by_symbol["BUY/USDT"]["action_status"] == "观察"
+    assert by_symbol["BUY/USDT"]["execution_status"] == "observe"
+    assert by_symbol["WAIT/USDT"]["action_status"] == "观察"
+    assert by_symbol["WAIT/USDT"]["execution_status"] == "observe"
+    assert by_symbol["OBS/USDT"]["action_status"] == "观察"
+    plan = json.loads(by_symbol["BUY/USDT"]["entry_plan_json"])
+    assert plan["market_risk_gate"]["blocked_action"] == "可即刻买入"
+    assert plan["market_risk_gate"]["final_action"] == "观察"

tests/test_paper_trading.py

@@ -14,6 +14,7 @@ from app.db.paper_trading import (
     list_paper_trades,
     reset_paper_trading_data,
     send_paper_trading_report,
+    sync_pending_paper_orders,
     sync_recommendation,
 )
 
@@ -675,6 +676,47 @@ def test_wait_pullback_paper_order_fills_when_price_touches(monkeypatch):
     assert order["fill_price"] == pytest.approx(95)
 
 
+def test_pending_paper_order_reconciles_from_latest_price_cache(monkeypatch):
+    monkeypatch.setenv("ALPHAX_PAPER_TRADING_ENABLED", "1")
+    monkeypatch.setenv("ALPHAX_PAPER_TRADE_NOTIONAL_USDT", "100")
+    monkeypatch.setenv("ALPHAX_PAPER_TRADE_FEE_RATE", "0")
+    monkeypatch.setenv("ALPHAX_PAPER_TRADE_SLIPPAGE_PCT", "0")
+    altcoin_db.init_db()
+    rec_id = altcoin_db.create_recommendation(
+        symbol="CACHEFILL/USDT",
+        rec_state="蓄力",
+        rec_score=22,
+        entry_price=95,
+        stop_loss=90,
+        tp1=105,
+        tp2=112,
+        signals=["等待回踩"],
+        entry_plan={"entry_action": "等回踩", "entry_price": 95, "stop_loss": 90, "tp1": 105, "risk_reward_ok": True, "rr1": 2.0},
+    )
+    rec = {
+        "id": rec_id,
+        "symbol": "CACHEFILL/USDT",
+        "execution_status": "wait_pullback",
+        "action_status": "等回踩",
+        "entry_price": 95,
+        "stop_loss": 90,
+        "tp1": 105,
+        "tp2": 112,
+        "entry_plan": {"entry_action": "等回踩", "entry_price": 95, "stop_loss": 90, "tp1": 105, "risk_reward_ok": True, "rr1": 2.0},
+    }
+    sync_recommendation(rec, 100, event_time="2026-05-16T10:00:00")
+    altcoin_db.update_latest_price_cache("CACHEFILL/USDT", 94.9, updated_at="2026-05-16T10:05:00", source="test")
+
+    result = sync_pending_paper_orders(event_time="2026-05-16T10:05:00")
+
+    assert result["filled_count"] == 1
+    assert result["results"][0]["paper_order"]["filled"] is True
+    trade = list_paper_trades()["items"][0]
+    assert trade["symbol"] == "CACHEFILL/USDT"
+    order = list_paper_orders(status="filled")["items"][0]
+    assert order["fill_price"] == pytest.approx(95)
+
+
 def test_wait_pullback_order_cancels_when_recommendation_invalid(monkeypatch):
     monkeypatch.setenv("ALPHAX_PAPER_TRADING_ENABLED", "1")
     altcoin_db.init_db()

tests/test_price_streamer.py

@@ -6,15 +6,29 @@ from app.db.runtime_config_db import set_config
 from app.services import price_streamer
 
 
-@pytest.fixture
+def _paper_config(**overrides):
-def buy_now_rec(monkeypatch):
+    cfg = {
-    set_config("system", "paper_trading", {
         "enabled": True,
         "trade_notional_usdt": 5000,
         "trade_leverage": 5,
         "fee_rate": 0,
         "slippage_pct": 0,
-    })
+        "entry_gate_enabled": False,
+        "entry_min_rr": 1.2,
+        "order_min_rr": 1.2,
+        "max_cumulative_leverage": 0,
+        "max_stop_loss_leverage_risk_pct": 0,
+        "max_account_drawdown_pause_pct": 0,
+        "pause_after_weak_entries": 0,
+        "global_risk_gate_enabled": False,
+    }
+    cfg.update(overrides)
+    return cfg
+
+
+@pytest.fixture
+def buy_now_rec(monkeypatch):
+    set_config("system", "paper_trading", _paper_config())
     set_config("system", "price_streamer", {
         "enabled": True,
         "update_latest_price_cache": True,
@@ -95,13 +109,7 @@ def test_price_streamer_tracks_open_paper_trade_without_active_rec(buy_now_rec):
 
 
 def test_price_streamer_fills_pending_paper_order():
-    set_config("system", "paper_trading", {
+    set_config("system", "paper_trading", _paper_config())
-        "enabled": True,
-        "trade_notional_usdt": 5000,
-        "trade_leverage": 5,
-        "fee_rate": 0,
-        "slippage_pct": 0,
-    })
     set_config("system", "price_streamer", {
         "enabled": True,
         "update_latest_price_cache": True,
@@ -140,13 +148,7 @@ def test_price_streamer_fills_pending_paper_order():
 
 
 def test_price_streamer_prioritizes_pending_order_over_same_symbol_recommendation():
-    set_config("system", "paper_trading", {
+    set_config("system", "paper_trading", _paper_config())
-        "enabled": True,
-        "trade_notional_usdt": 5000,
-        "trade_leverage": 5,
-        "fee_rate": 0,
-        "slippage_pct": 0,
-    })
     set_config("system", "price_streamer", {
         "enabled": True,
         "update_latest_price_cache": True,

tests/test_review_accuracy_pipeline.py

@@ -161,7 +161,7 @@ def test_daily_factor_weight_governance_promotes_and_eliminates(monkeypatch, tem
     monkeypatch.setattr(review_engine, "update_signal_weight", lambda signal, weight: changes.append((signal, weight)))
     monkeypatch.setattr(review_engine, "get_signal_weights", lambda: {
         "good_factor": {"category": "前瞻", "total_count": 12, "hit_rate": 70, "avg_pnl": 4.2, "weight": 1.0},
-        "bad_factor": {"category": "前瞻", "total_count": 12, "hit_rate": 5, "avg_pnl": -2.0, "weight": 1.2},
+        "bad_factor": {"category": "前瞻", "total_count": 20, "hit_rate": 5, "avg_pnl": -2.0, "weight": 1.2},
         "thin_factor": {"category": "PA", "total_count": 2, "hit_rate": 100, "avg_pnl": 6.0, "weight": 1.0},
     })