diff --git a/rqalpha/mod/rqalpha_mod_sys_analyser/mod.py b/rqalpha/mod/rqalpha_mod_sys_analyser/mod.py
index f483675e5..5398b4d53 100644
--- a/rqalpha/mod/rqalpha_mod_sys_analyser/mod.py
+++ b/rqalpha/mod/rqalpha_mod_sys_analyser/mod.py
@@ -116,21 +116,9 @@ def start_up(self, env, mod_config):
 
             self._plot_store = PlotStore(env)
             export_as_api(self._plot_store.plot)
-
-    def get_benchmark_daily_returns(self):
-        if self._benchmark is None:
-            return np.nan
-        daily_return_list = []
-        weights = 0
-        for benchmark in self._benchmark:
-            bar = self._env.data_proxy.get_bar(benchmark[0], self._env.calendar_dt, '1d')
-            if bar.close != bar.close:
-                daily_return_list.append((0.0, benchmark[1]))
-            else:
-                daily_return_list.append((bar.close / bar.prev_close - 1.0, benchmark[1]))
-            weights += benchmark[1]
-        return sum([daily[0] * daily[1] / weights for daily in daily_return_list])
     
+    NULL_OID = "null"
+
     def generate_benchmark_daily_returns_and_portfolio(self, event):
         _s = self._env.config.base.start_date
         _e = self._env.config.base.end_date
@@ -143,37 +131,41 @@ def generate_benchmark_daily_returns_and_portfolio(self, event):
         self._benchmark_daily_returns = np.zeros(len(trading_dates))
         weights = 0
         for order_book_id, weight in self._benchmark:
-            ins = self._env.data_proxy.instrument(order_book_id)
-            if ins is None:
-                raise RuntimeError(
-                    _("benchmark {} not exists, please entry correct order_book_id").format(order_book_id)
-                )
-            bars = self._env.data_proxy.history_bars(
-                order_book_id = order_book_id,
-                bar_count = len(trading_dates) + 1,  # Get an extra day for calculation
-                frequency = "1d",
-                field = ["datetime", "close"],
-                dt = _e,
-                skip_suspended=False,
-            )
-            if len(bars) == len(trading_dates) + 1:
-                if convert_int_to_date(bars[1]['datetime']).date() != _s:
-                    raise RuntimeError(_(
-                        "benchmark {} missing data between backtest start date {} and end date {}").format(order_book_id, _s, _e)
-                    )
-                daily_returns = (bars['close'] / np.roll(bars['close'], 1) - 1.0)[1: ]
-                self._benchmark_daily_returns = self._benchmark_daily_returns + daily_returns * weight
-                weights += weight
+            if order_book_id == self.NULL_OID:
+                daily_returns = np.zeros(len(trading_dates))
             else:
-                if len(bars) == 0:
-                    (available_s, available_e) = (ins.listed_date, ins.de_listed_date)
-                else:
-                    (available_s, available_e) = (convert_int_to_date(bars[0]['datetime']).date(), convert_int_to_date(bars[-1]['datetime']).date())
-                raise RuntimeError(
-                    _("benchmark {} available data start date {} >= backtest start date {} or end date {} <= backtest end "
-                    "date {}").format(order_book_id, available_s, _s, available_e, _e)
+                ins = self._env.data_proxy.instrument(order_book_id)
+                if ins is None:
+                    raise RuntimeError(
+                        _("benchmark {} not exists, please entry correct order_book_id").format(order_book_id)
+                    )
+                bars = self._env.data_proxy.history_bars(
+                    order_book_id = order_book_id,
+                    bar_count = len(trading_dates) + 1,  # Get an extra day for calculation
+                    frequency = "1d",
+                    field = ["datetime", "close"],
+                    dt = _e,
+                    skip_suspended=False,
                 )
-        self._benchmark_daily_returns = self._benchmark_daily_returns / weight
+                if len(bars) == len(trading_dates) + 1:
+                    if convert_int_to_date(bars[1]['datetime']).date() != _s:
+                        raise RuntimeError(_(
+                            "benchmark {} missing data between backtest start date {} and end date {}").format(order_book_id, _s, _e)
+                        )
+                    daily_returns = (bars['close'] / np.roll(bars['close'], 1) - 1.0)[1: ]
+                else:
+                    if len(bars) == 0:
+                        (available_s, available_e) = (ins.listed_date, ins.de_listed_date)
+                    else:
+                        (available_s, available_e) = (convert_int_to_date(bars[0]['datetime']).date(), convert_int_to_date(bars[-1]['datetime']).date())
+                    raise RuntimeError(
+                        _("benchmark {} available data start date {} >= backtest start date {} or end date {} <= backtest end "
+                        "date {}").format(order_book_id, available_s, _s, available_e, _e)
+                    )
+            self._benchmark_daily_returns = self._benchmark_daily_returns + daily_returns * weight
+            weights += weight
+
+        self._benchmark_daily_returns = self._benchmark_daily_returns / weights
         
         # generate benchmark portfolio
         unit_net_value = (self._benchmark_daily_returns + 1).cumprod()
@@ -229,7 +221,8 @@ def _parse_benchmark(benchmarks):
         benchmark_list = benchmarks.split(',')
         if len(benchmark_list) == 1:
             if len(benchmark_list[0].split(':')) > 1:
-                result.append((benchmark_list[0].split(':')[0], 1.0))
+                oid, weight = benchmark_list[0].split(':')
+                result.append((oid, float(weight)))
                 return result
             result.append((benchmark_list[0], 1.0))
             return result
@@ -360,7 +353,7 @@ def tear_down(self, code, exception=None):
                 summary["benchmark_symbol"] = self._env.data_proxy.instrument(benchmark_obid).symbol
             else:
                 summary["benchmark"] = ",".join(f"{o}:{w}" for o, w in self._benchmark)
-                summary["benchmark_symbol"] = ",".join(f"{self._env.data_proxy.instrument(o).symbol}:{w}" for o, w in self._benchmark)
+                summary["benchmark_symbol"] = ",".join(f"{self._env.data_proxy.instrument(o).symbol if o != self.NULL_OID else 'null'}:{w}" for o, w in self._benchmark)
 
         risk_free_rate = data_proxy.get_risk_free_rate(self._env.config.base.start_date, self._env.config.base.end_date)
         risk = Risk(
diff --git a/setup.cfg b/setup.cfg
index ed0fb1f0b..2acad2ec2 100644
--- a/setup.cfg
+++ b/setup.cfg
@@ -5,7 +5,7 @@
 
 [metadata]
 name = rqalpha
-version = 5.4.3
+version = 5.4.4
 
 [versioneer]
 VCS = git
diff --git a/tests/api_tests/mod/sys_analyser/test_negative_benchmark.py b/tests/api_tests/mod/sys_analyser/test_negative_benchmark.py
new file mode 100644
index 000000000..7afb02b5a
--- /dev/null
+++ b/tests/api_tests/mod/sys_analyser/test_negative_benchmark.py
@@ -0,0 +1,55 @@
+# -*- coding: utf-8 -*-
+# 版权所有 2019 深圳米筐科技有限公司（下称“米筐科技”）
+#
+# 除非遵守当前许可，否则不得使用本软件。
+#
+#     * 非商业用途（非商业用途指个人出于非商业目的使用本软件，或者高校、研究所等非营利机构出于教育、科研等目的使用本软件）：
+#         遵守 Apache License 2.0（下称“Apache 2.0 许可”），您可以在以下位置获得 Apache 2.0 许可的副本：http://www.apache.org/licenses/LICENSE-2.0。
+#         除非法律有要求或以书面形式达成协议，否则本软件分发时需保持当前许可“原样”不变，且不得附加任何条件。
+#
+#     * 商业用途（商业用途指个人出于任何商业目的使用本软件，或者法人或其他组织出于任何目的使用本软件）：
+#         未经米筐科技授权，任何个人不得出于任何商业目的使用本软件（包括但不限于向第三方提供、销售、出租、出借、转让本软件、本软件的衍生产品、引用或借鉴了本软件功能或源代码的产品或服务），任何法人或其他组织不得出于任何目的使用本软件，否则米筐科技有权追究相应的知识产权侵权责任。
+#         在此前提下，对本软件的使用同样需要遵守 Apache 2.0 许可，Apache 2.0 许可与本许可冲突之处，以本许可为准。
+#         详细的授权流程，请联系 public@ricequant.com 获取。
+
+import pandas as pd
+from numpy import isclose, array
+
+from rqalpha.apis import *
+
+
+__config__ = {
+    "base": {
+        "start_date": "2024-11-04",
+        "end_date": "2024-11-08",
+        "frequency": "1d",
+        "accounts": {
+            "stock": 10000000,
+        },
+    },
+    "extra": {
+        "log_level": "error",
+    },
+    "mod": {
+        "sys_analyser": {
+            "benchmark": "000300.XSHG:-1,null:2",
+        }
+    }
+}
+
+
+def test_negative_benchmark():
+    def handle_bar(context, _):
+        from rqalpha.environment import Environment
+
+        env = Environment.get_instance()
+        df = pd.DataFrame(env.mod_dict["sys_analyser"]._total_benchmark_portfolios)
+        df['date'] = pd.to_datetime(df['date'])
+        benchmark_portfolio = df.set_index('date').sort_index()        
+
+        assert isclose(
+            (benchmark_portfolio / benchmark_portfolio.shift(1, fill_value=1) - 1)["unit_net_value"].values, 
+            array([-0.01407232, -0.02530206,  0.00501645, -0.03016987,  0.01004613])
+        ).all()
+
+    return locals()
\ No newline at end of file
diff --git a/versioneer.py b/versioneer.py
index 645fc868a..3d5296740 100644
--- a/versioneer.py
+++ b/versioneer.py
@@ -1665,13 +1665,9 @@ def render_pep440_ricequant(pieces):
     # # 如果是dev和master分支或者hotfix分支来的，或者是一个tag，那就用pep440的版本号，否则带上git commit id
     if tracking_branch in ["origin/develop", "origin/master"] or tracking_branch.startswith("origin/hotfix/") or pieces[
         "distance"] == 0:
-        if pieces["dirty"]:
-            rendered += ".dirty"
         return rendered
 
     rendered += "+%s" % format(pieces["short"])
-    if pieces["dirty"]:
-        rendered += ".dirty"
     return rendered