完成回测的基本计算

2026-01-20 10:29:43 +08:00
parent 36db9d05ea
commit 04a3fbdf8e
5 changed files with 906 additions and 376 deletions
--- a/backtest.ipynb
+++ b/backtest.ipynb
--- a/note.md
+++ b/note.md
@@ -0,0 +1,126 @@
+# 基础时间与权益指标
+
+## Exposure Time [%] - 持仓时间占比（%）
+
+衡量策略在统计周期内实际持有头寸的时间占总时间的百分比。反映策略的“活跃度”（如占比100%=全程持仓，占比低=多数时间空仓），辅助判断交易频率和资金使用效率。
+
+## Equity Final [$] - 最终权益
+
+策略运行结束时的账户总资金（本金+盈亏），是最基础的收益结果指标，直接体现最终资金规模。
+
+## Equity Peak [$] - 权益峰值
+
+统计周期内账户权益的最大值。对比最终权益可判断策略是否从峰值回落，结合回撤指标能评估收益稳定性。
+
+# 收益率类指标（核心收益衡量）
+
+## Return [%] - 总收益率（%）
+
+周期内总收益占初始本金的百分比（`(最终权益-初始本金)/初始本金×100%`），直观反映整体盈利/亏损程度，但未考虑时间因素。
+
+## Buy & Hold Return [%] - 买入并持有收益率（%）
+
+作为“基准收益率”，模拟“买入标的后长期持有不操作”的收益。用于对比策略收益，判断策略是否跑赢简单持有，是评估策略有效性的核心参照。
+
+## Return (Ann.) [%] - 年化收益率（%）
+
+将周期总收益折算为按年计算的收益（`(1+总收益率)^(365/持仓天数) - 1`），消除时间周期差异，是跨周期对比不同策略收益的关键指标。
+
+## CAGR [%] - 复合年均增长率（%）
+
+即“年均复合收益率”（`(最终权益/初始本金)^(1/年数) - 1`），衡量资金长期复利增长速度，更贴合多年期策略的收益评估，避免短期高收益误导。
+
+# 风险类指标（衡量亏损与波动风险）
+
+## Volatility (Ann.) [%] - 年化波动率（%）
+
+收益波动程度的年化标准差，反映收益稳定性。波动率越高，收益起伏越大，策略“震荡风险”越高；波动率低则收益更平稳。
+
+## Max. Drawdown [%] - 最大回撤（%）
+
+权益从峰值回落至谷值的最大幅度（`(峰值-谷值)/峰值×100%`），核心风险指标。比如最大回撤20%，说明策略曾从最高点亏掉20%，直接体现抗风险能力。
+
+## Avg. Drawdown [%] - 平均回撤（%）
+
+所有回撤的平均幅度。若最大回撤远高于平均回撤，说明策略曾出现极端亏损；两者接近则回撤风险更均匀。
+
+## Max. Drawdown Duration - 最大回撤持续时长
+
+最大回撤从峰值回落至恢复峰值的时间（天/月），反映策略从最大亏损中回血的速度，时长越长则恢复能力越弱。
+
+## Avg. Drawdown Duration - 平均回撤持续时长
+
+所有回撤周期的平均时间，辅助判断策略亏损后的恢复效率。
+
+# 风险收益比指标（性价比核心）
+
+## Sharpe Ratio - 夏普比率
+
+衡量“每承担1单位风险的超额收益”（`(年化收益率-无风险利率)/年化波动率`
+），无风险利率通常取国债利率（3%左右）。比率越高性价比越高（>1较好，>2优秀）。
+
+## Sortino Ratio - 索提诺比率
+
+改进版夏普比率，仅用“下行波动率”（亏损方向波动）计算风险，更贴合投资者对“亏损风险”的关注，适合评估稳健型策略。
+
+## Calmar Ratio - 卡尔玛比率
+
+`年化收益率/最大回撤（绝对值）`，直接反映“收益能否覆盖最大亏损”，比率越高，策略用小回撤换高收益的能力越强。
+
+# 资产定价系数（市场相关性）
+
+## Alpha [%] - 阿尔法系数（%）
+
+衡量策略“超额收益”（超出市场基准的收益），Alpha为正说明策略能跑赢市场，反映主动管理能力；为负则跑输市场。
+
+## Beta - 贝塔系数
+
+衡量策略与市场基准的波动同步性。Beta=1=和市场波动一致；Beta>1=策略更激进（波动更大）；Beta<1=策略更稳健；Beta<0=与市场反向波动。
+
+# 交易行为与质量指标
+
+## Trades - 交易总次数
+
+统计周期内总交易笔数，反映策略交易频率（高频/低频），结合胜率可判断策略有效性。
+
+## Win Rate [%] - 胜率（%）
+
+盈利交易次数/总次数×100%，直观反映“赚钱交易的比例”，但需结合盈亏比（平均盈利/平均亏损）使用（如胜率50%+盈亏比3:1仍盈利）。
+
+## Best Trade [%] - 最佳单笔交易收益率（%）
+
+单次盈利幅度最大值，反映策略盈利上限，但需警惕是否由偶然高收益拉高整体表现。
+
+## Worst Trade [%] - 最差单笔交易收益率（%）
+
+单次亏损幅度最大值（通常为负），反映单笔交易的最大亏损风险。
+
+## Avg. Trade [%] - 平均单笔交易收益率（%）
+
+所有交易收益率的平均值，反映单笔交易的平均盈利能力，为正则单笔交易平均赚钱。
+
+## Max. Trade Duration - 单笔交易最长持有时长
+
+单次持仓的最长时间，定义策略风格（日内/短线/中长线）。
+
+## Avg. Trade Duration - 单笔交易平均持有时长
+
+所有交易持仓时长的平均值，辅助判断资金周转效率。
+
+# 综合绩效指标（策略可持续性）
+
+## Profit Factor - 盈利因子
+
+`总盈利金额/总亏损金额（绝对值）`，核心指标。>1说明总盈利>总亏损，数值越高盈利能力越强（如盈利因子2=每亏1元赚2元）。
+
+## Expectancy [%] - 期望收益（%）
+
+每笔交易的预期收益（`胜率×平均盈利 - (1-胜率)×平均亏损`），为正则策略长期可盈利，数值越高预期收益越好。
+
+## SQN - 系统质量数
+
+衡量策略收益的“统计显著性”（`√交易次数 × 期望收益 / 交易收益标准差`）。SQN>1.6合格，>2.5优秀，>3.0卓越，判断收益是否为偶然结果。
+
+## Kelly Criterion - 凯利准则
+
+计算“最优仓位比例”（`(胜率×(盈亏比+1)-1)/盈亏比`），确定每笔交易的最佳资金投入比例，最大化长期复利收益，同时控制破产风险（实际应用中通常减半使用）。
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -5,8 +5,10 @@ description = "Stock analysis"
 requires-python = ">=3.14"
 dependencies = [
    "backtesting~=0.6.5",
+    "duckdb>=1.4.3",
    "jupyter~=1.1.1",
    "matplotlib~=3.10.8",
+    "mplfinance>=0.12.10b0",
    "pandas~=2.3.3",
    "pandas-stubs~=2.3.3",
    "peewee~=3.19.0",
--- a/sql2dataframe.ipynb
+++ b/sql2dataframe.ipynb
@@ -0,0 +1,72 @@
+{
+ "cells": [
+  {
+   "metadata": {},
+   "cell_type": "code",
+   "outputs": [],
+   "execution_count": null,
+   "source": [
+    "import pandas as pd\n",
+    "\n",
+    "host = '81.71.3.24'\n",
+    "port = 6785\n",
+    "database = 'leopard_dev'\n",
+    "username = 'leopard'\n",
+    "password = '9NEzFzovnddf@PyEP?e*AYAWnCyd7UhYwQK$pJf>7?ccFiN^x4$eKEZ5~E<7<+~X'"
+   ],
+   "id": "1e8d815ee9b8c936"
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "initial_id",
+   "metadata": {
+    "collapsed": true
+   },
+   "outputs": [],
+   "source": [
+    "import psycopg2\n",
+    "\n",
+    "dailies_df = pd.DataFrame()\n",
+    "\n",
+    "with psycopg2.connect(host=host, port=port, database=database, user=username, password=password) as connection:\n",
+    "    with connection.cursor() as cursor:\n",
+    "        # language=PostgreSQL\n",
+    "        cursor.execute(\n",
+    "            \"\"\"select trade_date, open, close, high, low, factor\n",
+    "from leopard_daily daily\n",
+    "         left join leopard_stock stock on stock.id = daily.stock_id\n",
+    "where stock.code = '000001.SZ'\n",
+    "  and daily.trade_date between '2025-01-01 00:00:00' and '2025-12-31 23:59:59'\n",
+    "order by daily.trade_date\"\"\"\n",
+    "            )\n",
+    "        rows = cursor.fetchall()\n",
+    "\n",
+    "        dailies_df = pd.DataFrame.from_records(rows, columns=['trade_date', 'open', 'close', 'high', 'low', 'factor'])\n",
+    "\n",
+    "dailies_df"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 2
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython2",
+   "version": "2.7.6"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}
--- a/uv.lock
+++ b/uv.lock
@@ -338,6 +338,21 @@ wheels = [
    { url = "https://files.pythonhosted.org/packages/07/6c/aa3f2f849e01cb6a001cd8554a88d4c77c5c1a31c95bdf1cf9301e6d9ef4/defusedxml-0.7.1-py2.py3-none-any.whl", hash = "sha256:a352e7e428770286cc899e2542b6cdaedb2b4953ff269a210103ec58f6198a61", size = 25604, upload-time = "2021-03-08T10:59:24.45Z" },
 ]

+[[package]]
+name = "duckdb"
+version = "1.4.3"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/7f/da/17c3eb5458af69d54dedc8d18e4a32ceaa8ce4d4c699d45d6d8287e790c3/duckdb-1.4.3.tar.gz", hash = "sha256:fea43e03604c713e25a25211ada87d30cd2a044d8f27afab5deba26ac49e5268", size = 18478418, upload-time = "2025-12-09T10:59:22.945Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/b6/f4/a38651e478fa41eeb8e43a0a9c0d4cd8633adea856e3ac5ac95124b0fdbf/duckdb-1.4.3-cp314-cp314-macosx_10_15_universal2.whl", hash = "sha256:316711a9e852bcfe1ed6241a5f654983f67e909e290495f3562cccdf43be8180", size = 29042272, upload-time = "2025-12-09T10:58:51.826Z" },
+    { url = "https://files.pythonhosted.org/packages/16/de/2cf171a66098ce5aeeb7371511bd2b3d7b73a2090603b0b9df39f8aaf814/duckdb-1.4.3-cp314-cp314-macosx_10_15_x86_64.whl", hash = "sha256:9e625b2b4d52bafa1fd0ebdb0990c3961dac8bb00e30d327185de95b68202131", size = 15419343, upload-time = "2025-12-09T10:58:54.439Z" },
+    { url = "https://files.pythonhosted.org/packages/35/28/6b0a7830828d4e9a37420d87e80fe6171d2869a9d3d960bf5d7c3b8c7ee4/duckdb-1.4.3-cp314-cp314-macosx_11_0_arm64.whl", hash = "sha256:130c6760f6c573f9c9fe9aba56adba0fab48811a4871b7b8fd667318b4a3e8da", size = 13748905, upload-time = "2025-12-09T10:58:56.656Z" },
+    { url = "https://files.pythonhosted.org/packages/15/4d/778628e194d63967870873b9581c8a6b4626974aa4fbe09f32708a2d3d3a/duckdb-1.4.3-cp314-cp314-manylinux_2_26_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:20c88effaa557a11267706b01419c542fe42f893dee66e5a6daa5974ea2d4a46", size = 18487261, upload-time = "2025-12-09T10:58:58.866Z" },
+    { url = "https://files.pythonhosted.org/packages/c6/5f/87e43af2e4a0135f9675449563e7c2f9b6f1fe6a2d1691c96b091f3904dd/duckdb-1.4.3-cp314-cp314-manylinux_2_26_x86_64.manylinux_2_28_x86_64.whl", hash = "sha256:1b35491db98ccd11d151165497c084a9d29d3dc42fc80abea2715a6c861ca43d", size = 20497138, upload-time = "2025-12-09T10:59:01.241Z" },
+    { url = "https://files.pythonhosted.org/packages/94/41/abec537cc7c519121a2a83b9a6f180af8915fabb433777dc147744513e74/duckdb-1.4.3-cp314-cp314-win_amd64.whl", hash = "sha256:23b12854032c1a58d0452e2b212afa908d4ce64171862f3792ba9a596ba7c765", size = 12836056, upload-time = "2025-12-09T10:59:03.388Z" },
+    { url = "https://files.pythonhosted.org/packages/b1/5a/8af5b96ce5622b6168854f479ce846cf7fb589813dcc7d8724233c37ded3/duckdb-1.4.3-cp314-cp314-win_arm64.whl", hash = "sha256:90f241f25cffe7241bf9f376754a5845c74775e00e1c5731119dc88cd71e0cb2", size = 13527759, upload-time = "2025-12-09T10:59:05.496Z" },
+]
+
 [[package]]
 name = "executing"
 version = "2.2.1"
@@ -850,8 +865,10 @@ version = "0.1.0"
 source = { virtual = "." }
 dependencies = [
    { name = "backtesting" },
+    { name = "duckdb" },
    { name = "jupyter" },
    { name = "matplotlib" },
+    { name = "mplfinance" },
    { name = "pandas" },
    { name = "pandas-stubs" },
    { name = "peewee" },
@@ -861,8 +878,10 @@ dependencies = [
 [package.metadata]
 requires-dist = [
    { name = "backtesting", specifier = "~=0.6.5" },
+    { name = "duckdb", specifier = ">=1.4.3" },
    { name = "jupyter", specifier = "~=1.1.1" },
    { name = "matplotlib", specifier = "~=3.10.8" },
+    { name = "mplfinance", specifier = ">=0.12.10b0" },
    { name = "pandas", specifier = "~=2.3.3" },
    { name = "pandas-stubs", specifier = "~=2.3.3" },
    { name = "peewee", specifier = "~=3.19.0" },
@@ -953,6 +972,19 @@ wheels = [
    { url = "https://files.pythonhosted.org/packages/9b/f7/4a5e785ec9fbd65146a27b6b70b6cdc161a66f2024e4b04ac06a67f5578b/mistune-3.2.0-py3-none-any.whl", hash = "sha256:febdc629a3c78616b94393c6580551e0e34cc289987ec6c35ed3f4be42d0eee1", size = 53598, upload-time = "2025-12-23T11:36:33.211Z" },
 ]

+[[package]]
+name = "mplfinance"
+version = "0.12.10b0"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "matplotlib" },
+    { name = "pandas" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/6c/a9/34e7998d02fb58fae04f750444ce4e95e75f3a08dad17fb2d32098a97834/mplfinance-0.12.10b0.tar.gz", hash = "sha256:7da150b5851aa5119ad6e06b55e48338b619bb6773f1b85df5de67a5ffd917bf", size = 70117, upload-time = "2023-08-02T15:13:53.829Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/d7/d9/31c436ea7673c21a5bf3fc747bc7f63377582dfe845c3004d3e46f9deee0/mplfinance-0.12.10b0-py3-none-any.whl", hash = "sha256:76d3b095f05ff35de730751649de063bea4064d0c49b21b6182c82997a7f52bb", size = 75016, upload-time = "2023-08-02T15:13:52.022Z" },
+]
+
 [[package]]
 name = "narwhals"
 version = "2.15.0"