doc/samples/netvalueestimation.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "本例子是 v1.0 的基金净值预测引擎，现在已经升级的 v3.0，不需要在参考该 notebook 的代码，请直接使用 ``xa.QDIIPredict`` 类即可，\n",
    "各种情况都已做了妥善的封装和考虑"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import xalpha as xa\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 南方原油持仓及基准\n",
    "# 38335 WisdomTree WTI Crude Oil\n",
    "# 44793 United States 12 Month Oil Fund LP\n",
    "# 995771 UBS ETF CHCMCI Oil SF USD A-dis https://cn.investing.com/etfs/ubs-cmci-oil-sf-usd\n",
    "# 44794 United States Oil Fund LP\n",
    "# 44792 Invesco DB Oil Fund\n",
    "# 38324 WisdomTree Brent Crude Oil\n",
    "# 44634 United States Brent Oil Fund LP\n",
    "# 37450 WisdomTree Brent Crude Oil 1mth\n",
    "# 1014132 iPath S&P GSCI Crude Oil Total Return Index ETN\n",
    "\n",
    "# 8833 brent oil\n",
    "# 8849 wti oil\n",
    "\n",
    "#2019 第四季度季报披露\n",
    "# reference: http://pdf.dfcfw.com/pdf/H2_AN202001191374337197_1.pdf\n",
    "holdings_501018_19s3 = {\n",
    "    \"37450\": 15.31,\n",
    "    \"44792\": 13.64,\n",
    "    \"44794\": 13.46,\n",
    "    \"44634\": 12.48,\n",
    "    \"38335\": 11.73,\n",
    "    \"38324\": 9.45,\n",
    "    \"995771\": 6.53,\n",
    "    \"44793\": 6.02,\n",
    "    \"1014132\": 5.42,\n",
    "}\n",
    "#2019 第三季度季报披露\n",
    "# reference: http://pdf.dfcfw.com/pdf/H2_AN201910241369695857_1.pdf\n",
    "holdings_501018_19s4 = {\n",
    "    \"38335\": 7.34,\n",
    "    \"44793\": 8.14,\n",
    "    \"995771\": 8.68,\n",
    "    \"44794\": 9.63,\n",
    "    \"44792\": 11.6,\n",
    "    \"38324\": 15.04,\n",
    "    \"44634\": 15.42,\n",
    "    \"37450\": 17.51,\n",
    "    \"1014132\": 0.06,\n",
    "}\n",
    "holdings_501018_bc_cash = {\"8849\": 52.8, \"8833\": 35.2}\n",
    "# holdings_bc = {\"8849\": 60, \"8833\": 40}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "## 国泰商品相关基金持仓\n",
    "# 14218 PROSHARES ULTRA BLOOMBERG CRUDE OIL\n",
    "# 37471 Invesco DB US Dollar Index Bullish Fund (UUP)\n",
    "## 44792 Invesco DB Oil Fund (DBO)\n",
    "# 44798 VELOCITYSHARES 3X LONG CRUDE OIL ETN\n",
    "# 44718 INVESCO DB PRECIOUS METALS FUND\n",
    "# 9236 ISHARES SILVER TRUST (SLV)\n",
    "\n",
    "holdings_160216_19s4 = {\n",
    "    \"44794\": 16.9,\n",
    "    \"14218\": 14.03,\n",
    "    \"44634\": 11.77,\n",
    "    \"37471\": 10.98,\n",
    "    \"44792\": 9.1,\n",
    "    \"44798\": 7.82,\n",
    "    \"1014132\": 7.76,\n",
    "    \"44793\": 2.25,\n",
    "    \"44718\": 0.01,\n",
    "    \"9236\": 0,\n",
    "}\n",
    "# reference: http://pdf.dfcfw.com/pdf/H2_AN202001161374256079_1.pdf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 华宝油气\n",
    "#  38284 SPDR® S&P Oil & Gas Exploration & Production ETF (XOP)\n",
    "#  1010825 S&P Oil & Gas Exploration & Production Select Industry TR (SPSIOPTR)\n",
    "# 华宝油气直接持仓 SPSIOPTR 成分股，而非买入 XOP，直接用油气指数模拟持仓\n",
    "holdings_162411_19s4 = {\"1010825\": 90}\n",
    "# reference: http://pdf.dfcfw.com/pdf/H2_AN202001201374387272_1.pdf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def cache_data(data, *args, start=\"20180101\"):\n",
    "    for arg in args:\n",
    "        for k in arg:\n",
    "            if k not in data:\n",
    "                data[k] = xa.get_daily(k, start=start)\n",
    "\n",
    "\n",
    "def daily_increment(code, date, lastday=None, _default=None):\n",
    "    if not _default:\n",
    "        tds = xa.get_daily(code=code, end=date, prev=15)\n",
    "        tds = tds[tds[\"date\"] <= date]\n",
    "        if not lastday:\n",
    "            ratio = tds.iloc[-1][\"close\"] / tds.iloc[-2][\"close\"]\n",
    "        else:\n",
    "            tds2 = tds[tds[\"date\"] <= lastday]\n",
    "            ratio = tds.iloc[-1][\"close\"] / tds2.iloc[-1][\"close\"]\n",
    "        return ratio\n",
    "    # _default table dict is existing\n",
    "    tds = _default[code][_default[code][\"date\"] <= date]\n",
    "    if not lastday:\n",
    "        ratio = tds.iloc[-1][\"close\"] / tds.iloc[-2][\"close\"]\n",
    "    else:\n",
    "        tds2 = tds[tds[\"date\"] <= lastday]\n",
    "        ratio = tds.iloc[-1][\"close\"] / tds2.iloc[-1][\"close\"]\n",
    "    return ratio\n",
    "\n",
    "\n",
    "def evaluate_fluctuation(holding_dict, date, lastday=None, _default=None, _usd=\"\"):\n",
    "    price = 0\n",
    "    tot = 0\n",
    "    for k, v in holding_dict.items():\n",
    "        tot += v\n",
    "    remain = 100 - tot\n",
    "    for fundid, percent in holding_dict.items():\n",
    "        ratio = daily_increment(fundid, date, lastday, _default=_default)\n",
    "        #         print(fundid, ratio)\n",
    "        price += ratio * percent / 100\n",
    "    price += remain / 100  # currency part\n",
    "    if _usd:  # 美元资产需换算汇率, 中间价数据似乎精确度好于市场数据 2111\n",
    "        price = price * daily_increment(_usd, date, lastday, _default=_default)\n",
    "    return (price - 1) * 100\n",
    "\n",
    "def estimate_table(start, end, *cols, _default=None):\n",
    "    \"\"\"\n",
    "    \n",
    "    :param cols: Tuple[str, str, Dict]. (colname, _usd, holding_dict).\n",
    "    \"\"\"\n",
    "    compare_data = {\"date\": [],}\n",
    "    for col in cols:\n",
    "        compare_data[col[0]] = []\n",
    "    dl = pd.Series(pd.date_range(start=start, end=end))\n",
    "    dl = dl[dl.isin(xa.cons.opendate)]\n",
    "    for i, d in enumerate(dl):\n",
    "        if i == 0:\n",
    "            continue\n",
    "            \n",
    "        dstr = d.strftime(\"%Y%m%d\")\n",
    "        lstdstr = dl.iloc[i-1].strftime(\"%Y%m%d\")\n",
    "        compare_data[\"date\"].append(d)\n",
    "        for col in cols:\n",
    "            compare_data[col[0]].append( evaluate_fluctuation(col[2], dstr, lstdstr, _default=_default, _usd=col[1]) )\n",
    "    cpdf = pd.DataFrame(compare_data)\n",
    "    col0 = cols[0]\n",
    "    for col in cols[1:]:\n",
    "        cpdf[\"diff_\"+col0[0]+\"_\"+col[0]] = cpdf[col0[0]] - cpdf[col[0]]\n",
    "    return cpdf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "cache_table = {}\n",
    "cache_data(\n",
    "    cache_table,\n",
    "    holdings_160216_19s4,\n",
    "    holdings_162411_19s4,\n",
    "    holdings_501018_19s3,\n",
    "    holdings_501018_19s4,\n",
    "    holdings_501018_bc_cash,\n",
    "    [\"F501018\", \"USD/CNY\", \"2111\", \"F160216\", \"F162411\"]\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 南方原油净值预估"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [],
   "source": [
    "nfyydf = estimate_table(\n",
    "    \"20190101\",\n",
    "    \"20200305\",\n",
    "    (\"real\", \"\", {\"F501018\": 100}),\n",
    "    (\"fund_est\", \"USD/CNY\", holdings_501018_19s4),\n",
    "    (\"fund_est_2111\", \"2111\", holdings_501018_19s4),\n",
    "    (\"bc_est\", \"USD/CNY\", holdings_501018_bc_cash),\n",
    "    (\"bc_est_2111\", \"2111\", holdings_501018_bc_cash),\n",
    "    _default=cache_table,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>date</th>\n",
       "      <th>real</th>\n",
       "      <th>fund_est</th>\n",
       "      <th>fund_est_2111</th>\n",
       "      <th>bc_est</th>\n",
       "      <th>bc_est_2111</th>\n",
       "      <th>diff_real_fund_est</th>\n",
       "      <th>diff_real_fund_est_2111</th>\n",
       "      <th>diff_real_bc_est</th>\n",
       "      <th>diff_real_bc_est_2111</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>263</th>\n",
       "      <td>2020-02-07</td>\n",
       "      <td>-0.647948</td>\n",
       "      <td>-0.785591</td>\n",
       "      <td>-0.037261</td>\n",
       "      <td>-1.254779</td>\n",
       "      <td>-0.509988</td>\n",
       "      <td>0.137642</td>\n",
       "      <td>-0.610687</td>\n",
       "      <td>0.606830</td>\n",
       "      <td>-0.137960</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>264</th>\n",
       "      <td>2020-02-10</td>\n",
       "      <td>-1.096838</td>\n",
       "      <td>-1.426025</td>\n",
       "      <td>-1.800490</td>\n",
       "      <td>-1.428398</td>\n",
       "      <td>-1.802854</td>\n",
       "      <td>0.329187</td>\n",
       "      <td>0.703652</td>\n",
       "      <td>0.331560</td>\n",
       "      <td>0.706016</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>265</th>\n",
       "      <td>2020-02-11</td>\n",
       "      <td>0.429613</td>\n",
       "      <td>0.624201</td>\n",
       "      <td>0.310294</td>\n",
       "      <td>0.932186</td>\n",
       "      <td>0.617319</td>\n",
       "      <td>-0.194587</td>\n",
       "      <td>0.119319</td>\n",
       "      <td>-0.502573</td>\n",
       "      <td>-0.187706</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>266</th>\n",
       "      <td>2020-02-12</td>\n",
       "      <td>1.571827</td>\n",
       "      <td>2.244064</td>\n",
       "      <td>2.593395</td>\n",
       "      <td>2.198130</td>\n",
       "      <td>2.547303</td>\n",
       "      <td>-0.672238</td>\n",
       "      <td>-1.021568</td>\n",
       "      <td>-0.626303</td>\n",
       "      <td>-0.975477</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>267</th>\n",
       "      <td>2020-02-13</td>\n",
       "      <td>0.333007</td>\n",
       "      <td>0.504982</td>\n",
       "      <td>0.481938</td>\n",
       "      <td>0.701662</td>\n",
       "      <td>0.678573</td>\n",
       "      <td>-0.171975</td>\n",
       "      <td>-0.148931</td>\n",
       "      <td>-0.368655</td>\n",
       "      <td>-0.345566</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>268</th>\n",
       "      <td>2020-02-14</td>\n",
       "      <td>0.829754</td>\n",
       "      <td>1.198974</td>\n",
       "      <td>1.258412</td>\n",
       "      <td>1.343349</td>\n",
       "      <td>1.402872</td>\n",
       "      <td>-0.369220</td>\n",
       "      <td>-0.428658</td>\n",
       "      <td>-0.513595</td>\n",
       "      <td>-0.573118</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>269</th>\n",
       "      <td>2020-02-17</td>\n",
       "      <td>0.116178</td>\n",
       "      <td>-0.021618</td>\n",
       "      <td>-0.033048</td>\n",
       "      <td>0.298117</td>\n",
       "      <td>0.286651</td>\n",
       "      <td>0.137796</td>\n",
       "      <td>0.149225</td>\n",
       "      <td>-0.181939</td>\n",
       "      <td>-0.170473</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>270</th>\n",
       "      <td>2020-02-18</td>\n",
       "      <td>-0.232086</td>\n",
       "      <td>-0.002015</td>\n",
       "      <td>0.181235</td>\n",
       "      <td>-0.058525</td>\n",
       "      <td>0.124622</td>\n",
       "      <td>-0.230071</td>\n",
       "      <td>-0.413321</td>\n",
       "      <td>-0.173561</td>\n",
       "      <td>-0.356708</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>271</th>\n",
       "      <td>2020-02-19</td>\n",
       "      <td>2.268101</td>\n",
       "      <td>2.533318</td>\n",
       "      <td>2.275533</td>\n",
       "      <td>2.364867</td>\n",
       "      <td>2.107506</td>\n",
       "      <td>-0.265216</td>\n",
       "      <td>-0.007432</td>\n",
       "      <td>-0.096765</td>\n",
       "      <td>0.160596</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>272</th>\n",
       "      <td>2020-02-20</td>\n",
       "      <td>0.710833</td>\n",
       "      <td>0.478790</td>\n",
       "      <td>0.813269</td>\n",
       "      <td>0.618737</td>\n",
       "      <td>0.953681</td>\n",
       "      <td>0.232043</td>\n",
       "      <td>-0.102436</td>\n",
       "      <td>0.092097</td>\n",
       "      <td>-0.242848</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>273</th>\n",
       "      <td>2020-02-21</td>\n",
       "      <td>-1.138716</td>\n",
       "      <td>-1.355870</td>\n",
       "      <td>-1.555549</td>\n",
       "      <td>-0.612975</td>\n",
       "      <td>-0.814157</td>\n",
       "      <td>0.217154</td>\n",
       "      <td>0.416833</td>\n",
       "      <td>-0.525741</td>\n",
       "      <td>-0.324559</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>274</th>\n",
       "      <td>2020-02-24</td>\n",
       "      <td>-2.760590</td>\n",
       "      <td>-3.476302</td>\n",
       "      <td>-3.494192</td>\n",
       "      <td>-3.202966</td>\n",
       "      <td>-3.220907</td>\n",
       "      <td>0.715711</td>\n",
       "      <td>0.733602</td>\n",
       "      <td>0.442376</td>\n",
       "      <td>0.460317</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>275</th>\n",
       "      <td>2020-02-25</td>\n",
       "      <td>-2.085169</td>\n",
       "      <td>-1.550823</td>\n",
       "      <td>-1.753927</td>\n",
       "      <td>-2.434255</td>\n",
       "      <td>-2.635537</td>\n",
       "      <td>-0.534346</td>\n",
       "      <td>-0.331241</td>\n",
       "      <td>0.349086</td>\n",
       "      <td>0.550368</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>276</th>\n",
       "      <td>2020-02-26</td>\n",
       "      <td>-1.759648</td>\n",
       "      <td>-1.753248</td>\n",
       "      <td>-1.478477</td>\n",
       "      <td>-2.359271</td>\n",
       "      <td>-2.086195</td>\n",
       "      <td>-0.006400</td>\n",
       "      <td>-0.281171</td>\n",
       "      <td>0.599623</td>\n",
       "      <td>0.326547</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>277</th>\n",
       "      <td>2020-02-27</td>\n",
       "      <td>-4.152249</td>\n",
       "      <td>-3.919189</td>\n",
       "      <td>-4.286936</td>\n",
       "      <td>-2.476869</td>\n",
       "      <td>-2.850136</td>\n",
       "      <td>-0.233060</td>\n",
       "      <td>0.134687</td>\n",
       "      <td>-1.675381</td>\n",
       "      <td>-1.302113</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>278</th>\n",
       "      <td>2020-02-28</td>\n",
       "      <td>-3.291569</td>\n",
       "      <td>-3.027993</td>\n",
       "      <td>-2.995196</td>\n",
       "      <td>-3.936630</td>\n",
       "      <td>-3.904141</td>\n",
       "      <td>-0.263577</td>\n",
       "      <td>-0.296373</td>\n",
       "      <td>0.645061</td>\n",
       "      <td>0.612572</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>279</th>\n",
       "      <td>2020-03-02</td>\n",
       "      <td>2.964427</td>\n",
       "      <td>3.633032</td>\n",
       "      <td>3.550425</td>\n",
       "      <td>2.932988</td>\n",
       "      <td>2.850939</td>\n",
       "      <td>-0.668605</td>\n",
       "      <td>-0.585998</td>\n",
       "      <td>0.031439</td>\n",
       "      <td>0.113488</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>280</th>\n",
       "      <td>2020-03-03</td>\n",
       "      <td>0.309234</td>\n",
       "      <td>-0.041469</td>\n",
       "      <td>0.397138</td>\n",
       "      <td>0.034011</td>\n",
       "      <td>0.472950</td>\n",
       "      <td>0.350703</td>\n",
       "      <td>-0.087904</td>\n",
       "      <td>0.275223</td>\n",
       "      <td>-0.163715</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>281</th>\n",
       "      <td>2020-03-04</td>\n",
       "      <td>0.212608</td>\n",
       "      <td>-0.466318</td>\n",
       "      <td>-0.975291</td>\n",
       "      <td>-0.945985</td>\n",
       "      <td>-1.452505</td>\n",
       "      <td>0.678926</td>\n",
       "      <td>1.187899</td>\n",
       "      <td>1.158593</td>\n",
       "      <td>1.665113</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>282</th>\n",
       "      <td>2020-03-05</td>\n",
       "      <td>-1.665429</td>\n",
       "      <td>-1.922691</td>\n",
       "      <td>-1.608400</td>\n",
       "      <td>-1.934909</td>\n",
       "      <td>-1.620656</td>\n",
       "      <td>0.257262</td>\n",
       "      <td>-0.057029</td>\n",
       "      <td>0.269480</td>\n",
       "      <td>-0.044773</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          date      real  fund_est  fund_est_2111    bc_est  bc_est_2111  \\\n",
       "263 2020-02-07 -0.647948 -0.785591      -0.037261 -1.254779    -0.509988   \n",
       "264 2020-02-10 -1.096838 -1.426025      -1.800490 -1.428398    -1.802854   \n",
       "265 2020-02-11  0.429613  0.624201       0.310294  0.932186     0.617319   \n",
       "266 2020-02-12  1.571827  2.244064       2.593395  2.198130     2.547303   \n",
       "267 2020-02-13  0.333007  0.504982       0.481938  0.701662     0.678573   \n",
       "268 2020-02-14  0.829754  1.198974       1.258412  1.343349     1.402872   \n",
       "269 2020-02-17  0.116178 -0.021618      -0.033048  0.298117     0.286651   \n",
       "270 2020-02-18 -0.232086 -0.002015       0.181235 -0.058525     0.124622   \n",
       "271 2020-02-19  2.268101  2.533318       2.275533  2.364867     2.107506   \n",
       "272 2020-02-20  0.710833  0.478790       0.813269  0.618737     0.953681   \n",
       "273 2020-02-21 -1.138716 -1.355870      -1.555549 -0.612975    -0.814157   \n",
       "274 2020-02-24 -2.760590 -3.476302      -3.494192 -3.202966    -3.220907   \n",
       "275 2020-02-25 -2.085169 -1.550823      -1.753927 -2.434255    -2.635537   \n",
       "276 2020-02-26 -1.759648 -1.753248      -1.478477 -2.359271    -2.086195   \n",
       "277 2020-02-27 -4.152249 -3.919189      -4.286936 -2.476869    -2.850136   \n",
       "278 2020-02-28 -3.291569 -3.027993      -2.995196 -3.936630    -3.904141   \n",
       "279 2020-03-02  2.964427  3.633032       3.550425  2.932988     2.850939   \n",
       "280 2020-03-03  0.309234 -0.041469       0.397138  0.034011     0.472950   \n",
       "281 2020-03-04  0.212608 -0.466318      -0.975291 -0.945985    -1.452505   \n",
       "282 2020-03-05 -1.665429 -1.922691      -1.608400 -1.934909    -1.620656   \n",
       "\n",
       "     diff_real_fund_est  diff_real_fund_est_2111  diff_real_bc_est  \\\n",
       "263            0.137642                -0.610687          0.606830   \n",
       "264            0.329187                 0.703652          0.331560   \n",
       "265           -0.194587                 0.119319         -0.502573   \n",
       "266           -0.672238                -1.021568         -0.626303   \n",
       "267           -0.171975                -0.148931         -0.368655   \n",
       "268           -0.369220                -0.428658         -0.513595   \n",
       "269            0.137796                 0.149225         -0.181939   \n",
       "270           -0.230071                -0.413321         -0.173561   \n",
       "271           -0.265216                -0.007432         -0.096765   \n",
       "272            0.232043                -0.102436          0.092097   \n",
       "273            0.217154                 0.416833         -0.525741   \n",
       "274            0.715711                 0.733602          0.442376   \n",
       "275           -0.534346                -0.331241          0.349086   \n",
       "276           -0.006400                -0.281171          0.599623   \n",
       "277           -0.233060                 0.134687         -1.675381   \n",
       "278           -0.263577                -0.296373          0.645061   \n",
       "279           -0.668605                -0.585998          0.031439   \n",
       "280            0.350703                -0.087904          0.275223   \n",
       "281            0.678926                 1.187899          1.158593   \n",
       "282            0.257262                -0.057029          0.269480   \n",
       "\n",
       "     diff_real_bc_est_2111  \n",
       "263              -0.137960  \n",
       "264               0.706016  \n",
       "265              -0.187706  \n",
       "266              -0.975477  \n",
       "267              -0.345566  \n",
       "268              -0.573118  \n",
       "269              -0.170473  \n",
       "270              -0.356708  \n",
       "271               0.160596  \n",
       "272              -0.242848  \n",
       "273              -0.324559  \n",
       "274               0.460317  \n",
       "275               0.550368  \n",
       "276               0.326547  \n",
       "277              -1.302113  \n",
       "278               0.612572  \n",
       "279               0.113488  \n",
       "280              -0.163715  \n",
       "281               1.665113  \n",
       "282              -0.044773  "
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nfyydf.iloc[-20:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.13494830952722817,\n",
       " 0.2729164993400751,\n",
       " 0.422895067226587,\n",
       " 0.4503039730692361)"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(\n",
    "    np.mean(np.abs(nfyydf[\"diff_real_fund_est\"][-100:])),\n",
    "    np.mean(np.abs(nfyydf[\"diff_real_fund_est_2111\"][-100:])),\n",
    "    np.mean(np.abs(nfyydf[\"diff_real_bc_est\"][-100:])),\n",
    "    np.mean(np.abs(nfyydf[\"diff_real_bc_est_2111\"][-100:])),\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "可以看出，人民币中间价做汇率换算，比人民币市场价格要准很多，利用基金持仓的涨跌直接估算，也要比用基准准很多。\n",
    "南方原油大家一直抱怨不好估值，但我想 0.13 个百分点的平均误差水平应该是全网最准了。因为我想不到比直接从持仓开始算更准的方式了。\n",
    "如果有方式用最近100交易日的数据可以打败这一偏移水平，欢迎告诉我。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x11c2477b8>"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAAELCAYAAADX3k30AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4yLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvNQv5yAAAIABJREFUeJzt3Xl83XWd7/HX55zs+9o2bdKmpYXS\nCm2hYAEXVBhBGXBDUEcW9YHjdRnvjDN3HO91mzujjo5eFZfholcYBBd0FBHHQQQBtS1taQttKN1o\nmzZt0mbfc8753D/OSQ0haU6SsyQn7+fjcR45y+/8fp/8krzzO9/f9/f9mrsjIiKZK5DuAkREJLkU\n9CIiGU5BLyKS4RT0IiIZTkEvIpLhFPQiIhlOQS8ikuEU9CIiGU5BLyKS4bLSteGqqiqvr69P1+ZF\nRGalrVu3nnT36sm8J21BX19fz5YtW9K1eRGRWcnMDk32PWq6ERHJcAp6EZEMp6AXEclwCnoRkQyn\noBcRyXAKehGRDKegFxHJcAp6EZEMp6AXEclwabsyVkT+5N5Nh8d97Z0vX5zCSiQT6YheRCTDKehF\nRDLchEFvZnlmttnMdpjZLjP7zBjL3GJmLWa2PXZ7X3LKFRGRyYqnjX4AeK27d5tZNvCkmf3K3TeO\nWu6H7v6hxJcokhnO1A4vkkwTBr27O9Ade5gdu3kyixIRkcSJq43ezIJmth1oBh52901jLPZWM9tp\nZvebWV1CqxQRkSmLK+jdPezua4Fa4GIze9moRX4B1Lv7+cDDwF1jrcfMbjOzLWa2paWlZTp1i4hI\nnCbV68bd24FHgatGPX/K3QdiD+8ELhzn/Xe4+3p3X19dPamZsEREZIri6XVTbWZlsfv5wJXAc6OW\nqRnx8FqgIZFFiojI1MXT66YGuMvMgkT/MfzI3R80s88CW9z9AeAjZnYtEAJagVuSVbCIiExOPL1u\ndgLrxnj+kyPufxz4eGJLExGRRNCVsSIiGU5BLyKS4RT0IiIZTkEvIpLhFPQiIhlOQS8ikuEU9CIi\nGU5BLyKS4RT0IiIZTkEvIpLhFPQiIhlOQS8ikuEU9CIiGU5BLyKS4RT0IiIZTkEvIpLhFPQiIhlO\nQS8ikuEU9CIiGU5BLyKS4SYMejPLM7PNZrbDzHaZ2WfGWCbXzH5oZvvMbJOZ1SejWBERmbx4jugH\ngNe6+xpgLXCVmW0Ytcx7gTZ3Xw58BfhCYssUEZGpmjDoPao79jA7dvNRi10H3BW7fz/wOjOzhFUp\nIiJTFlcbvZkFzWw70Aw87O6bRi2yCDgC4O4hoAOoTGShIiIyNXEFvbuH3X0tUAtcbGYvm8rGzOw2\nM9tiZltaWlqmsgoREZmkSfW6cfd24FHgqlEvHQXqAMwsCygFTo3x/jvcfb27r6+urp5axSIiMinx\n9LqpNrOy2P184ErguVGLPQDcHLv/NuC37j66HV9ERNIgK45laoC7zCxI9B/Dj9z9QTP7LLDF3R8A\nvgP8u5ntA1qBG5NWsYiITMqEQe/uO4F1Yzz/yRH3+4HrE1uaiIgkgq6MFRHJcAp6EZEMp6AXEclw\nCnoRkQynoBdJs1A4wq5jHXT0DaW7FMlQ8XSvFJEk6OofYtPBVjYfbKV7IMSKeUXcetnSdJclGUhB\nL5IGA0NhvvbIXnoGw5wzv5iivCy2Hmqjsa2X2vKCdJcnGUZNNyJpsLupk57BMLdcWs/Nl9ZzzXk1\n5GcHeXSPxoCSxFPQi6TB9iPtlBVks3xeEQC52UEuPauShqZOjnf0p7k6yTQKepEU6+ofYn9LN2tq\nywiMmLbhkrMqyckK8NjzzWmsTjKRgl4kxZ492kHEYU1d2YueL8jJYsPSSp5p7OBk90CaqpNMpKAX\nSbEdjR0sKMljQUneS167bHklwYDxh/0n01CZZCoFvUgKtfYMcri19yVH88OK86Lt9vuae1JcmWQy\nBb1ICu1obAfg/NrScZdZWlXIye4Buvp1AZUkhoJeJEXcne1H2llSWUB5Qc64y9VXFgLwwqneVJUm\nGU5BL5IirT2DtHQNcP6i8Y/mARaW5ZMTDHDwpJpvJDEU9CIpcqQteoReX1V4xuWCAWNxZQEvKOgl\nQRT0IilyuLWPnGCA+WP0thltaVUhJzr76R0MpaAyyXQKepEUaWzrZVF5/osukhpPfWUhDhxSO70k\ngIJeJAX6h8Ica+9jcUV8A5bVlueTFTC100tCTBj0ZlZnZo+a2W4z22VmfzXGMpebWYeZbY/dPjnW\nukTmql3HolfD1sU5MmV2MEBteQEvnFLQy/TFM0xxCPgbd99mZsXAVjN72N13j1ruCXe/JvElisx+\nTx+O9p+vq8iP+z1Lqwr53fPNdA+EKMrViOIydRMe0bt7k7tvi93vAhqARckuTCSTPB0brbI4Lzvu\n9yytKiTisPVQWxIrk7lgUm30ZlYPrAM2jfHyJWa2w8x+ZWarx3n/bWa2xcy2tLRo3G2ZO7Yfbo+7\n2WbY4ooCAgabD55KUlUyV8Qd9GZWBPwE+Ki7d456eRuwxN3XAF8HfjbWOtz9Dndf7+7rq6urp1qz\nyKxyorOfo+191MV5InZYTlaAhWX5OqKXaYsr6M0sm2jIf9/dfzr6dXfvdPfu2P2HgGwzq0popSKz\n1HD7/OLy+Nvnhy0sy2f3sU7cPdFlyRwST68bA74DNLj7l8dZZkFsOczs4th69XlThOhsUtlBo6Zs\n8kFfU5pHZ3+Io+19SahM5op4TuVfBrwbeMbMtsee+wdgMYC7fxt4G/ABMwsBfcCNrkMQEQCePtzG\nqpoSsoOTv2xlYWn0n8OuY52aNFymbMKgd/cngTNeyufutwO3J6ookUwRCkfY2djBDRfVTen980vy\nCBjsPtbJ61cvSHB1MlfoyliRJNrb3E3fUJi140w0MpGcrADLqovY3TS6/4NI/BT0Ikm0+1g0oFcv\nLJnyOlbVlJxej8hUKOhFkqihqZOcrABLJxia+ExWLSzhaHsfHb2acUqmRkEvkkQNxzs5Z34xWVM4\nETtsVU3008Cupo5ElSVzjIJeJEncnYamrtNBPVWrYs0+ar6RqVLQiyRJc9cArT2DnFtTPK31VBXl\nMr8kVydkZcoU9CJJMnwEfu40j+hBJ2RlehT0IkkyfAS+MhFBv7CEfc3dDITC016XzD0KepEkaWjq\npLY8n9L8+IcmHs/qhaWEIs7eE90JqEzmGgW9SJI0NHUmpNkG/tTzRs03MhUKepEk6B8Kc/BkT8KC\nfnFFAYU5QZ2QlSlR0IskwZ7jXUQcVk2zx82wQMA4t6aEXcfUl14mT0EvkgTDR96rakoTts6VNcU8\nd7xLY9PLpCnoRZKgoamTotwsaqcw2ch4Vi4ooas/xLGO/oStU+YGBb1IEjQ0dbJyQTGBwBlH+J6U\nlQuizUB7jqudXiZHQS+SYO7Oc01dCTsRO+zsWNA3NHUldL2S+RT0IgnW2NZH10Do9Bg1iVKSl82i\nsnz2HFfQy+Qo6EUSbPhEbKKP6KPrLOY5Nd3IJCnoRRJs97FOAgbnzE9M18qRzllQzIGWHg2FIJMy\nYdCbWZ2ZPWpmu81sl5n91RjLmJl9zcz2mdlOM7sgOeWKzHy7mzpZVl1Efk4w4eteuaCEUMTZ39yT\n8HVL5orniD4E/I27rwI2AB80s1WjlrkaWBG73QZ8K6FViswiu491TnsM+vGc7nlzQs03Er8Jg97d\nm9x9W+x+F9AALBq12HXA3R61ESgzs5qEVysyw3X0DnG0vS8p7fMA9VWF5AQDPKeeNzIJk2qjN7N6\nYB2wadRLi4AjIx438tJ/BiIZ7/QVsQnucTMsOxhg+bwinlPPG5mEuIPezIqAnwAfdfcpfW40s9vM\nbIuZbWlpaZnKKkRmtD8NfZCcoIdo84163shkxBX0ZpZNNOS/7+4/HWORo0DdiMe1sedexN3vcPf1\n7r6+urp6KvWKzGgNTZ1UF+dSXZybtG2srCnmROcAbT2DSduGZJZ4et0Y8B2gwd2/PM5iDwA3xXrf\nbAA63L0pgXWKzArJPBE77JwF0fWr+UbilRXHMpcB7waeMbPtsef+AVgM4O7fBh4C3gDsA3qBWxNf\nqsjMNhiKsLe5i1edndxPq+eOGPPmkrMqk7otyQwTBr27PwmccWQmj46b+sFEFSUyG+1r7mYo7Ek7\nETusujiX8oJsHdFL3HRlrEiCpOJELICZsXJBCQ0KeomTgl4kQXYf6yQvO8DSqsKkb2tlTTHPH+8i\nHNEkJDIxBb1Iguxu6mDlghKCCRyDfjyrakroGwrzwikNhSATU9CLJIC705CEMejHs3phdIrCXcfU\nn14mpqAXSYBjHf109A0l/UTssOXzisgOGrsV9BKHeLpXisgEdh3tAGBVTeKHJr530+Exn68qyj19\nAljkTHREL5IAOxrbCQbsdJNKKtSU5uuIXuKioBdJgO1H2lm5oJi87MSPQT+emtI8TnYP0NzVn7Jt\nyuykoBeZpkjE2Xmkg7V1ZSndbk1ZHoCO6mVCaqMXmab9Ld10DYRYW1c2bnt6MtSU5APRC7UuP2de\nyrYrs4+O6EWm6ekj7QCsW5zaI/r8nCC15Wqnl4kp6EWmafuRdorzslhWVZTyba+qKVHPG5mQgl5k\nmnYcaWdNbRmBFFwRO9qqhSUcPNlD72Ao5duW2UNBLzINfYNhnjvelfITscNW1ZTgrrHp5cwU9CLT\n8OyxDsIRT1/Qx67EVTu9nImCXmQath+Onohdk6agX1SWT0leltrp5YwU9CLTsP1IO4vK8pM6R+yZ\nmBmrFpboiF7OSEEvMg3bj7SzNsXdKkdbvbCUhqZOhsKRtNYhM5eCXmSKmrv6Odrex7o0NdsMW1NX\nxkAowh6dkJVxKOhFpmjHkeiIlek6ETtsbW10+zsa29Nah8xcEwa9mX3XzJrN7NlxXr/czDrMbHvs\n9snElyky82w8cIqcrAAvW5S6ESvHUleRT0VhzukTwyKjxTPWzfeA24G7z7DME+5+TUIqEpklntx7\nkovrK1I6YuVYzIw1taU6opdxTXhE7+6PA60pqEVk1mju6mfPiS4uW16V7lKAaDv93uZuugd0hay8\nVKLa6C8xsx1m9iszWz3eQmZ2m5ltMbMtLS0tCdq0SOr9Yd8pAF65YmYE/dq6Mtxhp47qZQyJCPpt\nwBJ3XwN8HfjZeAu6+x3uvt7d11dXVydg0yLp8cTek5QVZLMqRZOBT2TN8AnZ2AlikZGmPR69u3eO\nuP+QmX3TzKrc/eR01y0yE7k7v993ksvOqkrLQGYjjRz/vqIwhwd3HqM0PxuAd758cbrKkhlm2kf0\nZrbAzCx2/+LYOk9Nd70iM9X+lm6Od/bzihnSbDOsrjyfI6296S5DZqAJj+jN7D7gcqDKzBqBTwHZ\nAO7+beBtwAfMLAT0ATe6uyetYpE0e3Jv9MPqK2bIidhhteUF7GjsoKNv6PRRvQjEEfTu/o4JXr+d\naPdLkTnhyX2nWFxRQF1FQbpLeZHhehrbeinNT2/ffplZdGWsyCQMhSNsPHBqxjXbANSU5hE0o7Gt\nL92lyAyjoBeZhJ2N7XQPhGZcsw1AdjDAgtI8jrSpnV5eTEEvMgmP7WkhYHDJssp0lzKm2vJ8jrb1\nEdFpMhlBQS8SJ3fnwZ1NbFhWSXlhTrrLGdOSykIGQhGaOvrTXYrMIAp6kTjtOtbJwZM9XHP+wnSX\nMq6lVYUAHGzpTnMlMpMo6EXi9ODOJoIB46qXLUh3KeMqzc+msjCHgyd70l2KzCAKepE4RJttjnHZ\n8ioqZmizzbClVYUcPNVDOKJ2eolS0IvEYUdjB41tfVxzfk26S5nQsupC+ociNGjCcIlR0IvE4cEd\nx8gOGq9fPXObbYYtrSoCYNNBjS4uUQp6kQlEItHeNq8+u3pWDC1Qmp9NRWEOGw9oyCmJUtCLTGDr\n4TaOd/bP6N42oy2rKmTzwVYiaqcXFPQiE/rZ00fJzQpwxar56S4lbkurCunoG+K5413pLkVmgGmP\nRy+SiYbHeR8YCvPjrY2srinhge3H0lxV/Ib70288cIpVC2fG5CiSPjqiFzmD7Y3tDIYivHyGDnkw\nnrKCHBZXFKidXgAFvci43J3NB1upKc2jrjw/3eVM2oZlFWxSO72goBcZ15G2Ppo6+rl4aQWxSdRm\nlUvPqqKjb4hnjmoe2blOQS8yjs0HT5GTFWBtbOLt2ebVZ1cTMHik4US6S5E0U9CLjKF3MMTOxg7W\n1ZWRmx1MdzlTUl6Yw4VLyvlNQ3O6S5E0U9CLjGHboTZCEefipRXpLmVaXnfufHY3ddLUoVmn5rIJ\ng97MvmtmzWb27Divm5l9zcz2mdlOM7sg8WWKpE444mw82MriigJqSmffSdiRrjh3HgCP6Kh+Toun\nH/33iE7+ffc4r18NrIjdXg58K/ZVZEYb7is/WkNTJ609g/zZLLpAajxnVRexuKKARxpO8BcblqS7\nHEmTCY/o3f1x4EyjI10H3O1RG4EyM5v5Q/yJjOMP+09Smp/N6oWl6S5l2syM1507j9/vP0XvYCjd\n5UiaJKKNfhFwZMTjxthzIrPO8c5+9rf0sGFpBcHA7OtSOZYrzp3PYCjC7/fp4qm5KqUnY83sNjPb\nYmZbWlpaUrlpkbj8cf9JsgLGRfWz+yTsSBfVV1Ccm6VulnNYIoL+KFA34nFt7LmXcPc73H29u6+v\nrq5OwKZFEqd3IMTTh9tZW1dGQW7mDAOVkxXgVWdX89vnmnWV7ByViKB/ALgp1vtmA9Dh7k0JWK9I\nSj0V61J56VlV6S4l4a5YNY/mrgGePtKW7lIkDSY8bDGz+4DLgSozawQ+BWQDuPu3gYeANwD7gF7g\n1mQVK5IsEXc2HTjFsqpCFpTmpbuchLty1QLys5/l/q1HuXBJ5jRLSXwmDHp3f8cErzvwwYRVJJIG\nzzV10d43xBvOy8wOY0W5WVz9sgU8uPMYn/rzVeTN0qt9ZWp0ZawIsPHgKUrzszm3JnPHbn/rhbV0\n9Yd4eLdOys41CnqZ81q6BtjX3M1F9ZnTpXIslyyrZGFpHvdvbUx3KZJiCnqZ8zYePEXQjIvqy9Nd\nSlIFAsZbLqjlib0tnOjsT3c5kkIKepnTBkJhth1qY/WiEorzstNdTtK95YJFRDw6D67MHQp6mdO2\nH2lnIBThklk2VeBULasu4oLFZdy/tZFoPwqZCxT0Mme5O5sORKcKXFxRkO5yUuatF9ayt7mbbYfb\n012KpIiCXuasI219HO+cvVMFTtWb1i6irCCbbz22P92lSIoo6GXO2nywlZxggDWzdKrAqSrMzeLW\nS5fym4YTPHe8M93lSApkzoAeIpMQnTQ7Oq5Npl48NN54+wA3X7qEOx7fzzce3c/X37EuhVVJOuiI\nXuak/9jWyFDYubh+bpyEHa2sIIe/uGQJv9x5jIMne9JdjiSZgl7mHHfn3s2HWVSWz6Ly2T1V4HS8\n7xXLyA4G+NZj+9JdiiSZgl7mnK2H2nj+RDcXZ9CY81NRXZzLjRfV8dNtR2ls6013OZJECnqZc+7d\ndJii3CzOr5v9UwVO1/tffRZZQeMzv9id7lIkiRT0Mqe09Qzy4DNNXLd2IblZmXkSdjIWluXz11ee\nzcO7T/Cfzx5PdzmSJAp6mVPu39rIYCjCX2xYku5SZoz3XLaUVTUlfOqBZ+nsH0p3OZIECnqZMyIR\n5/ubDrF+SXlGD0c8WVnBAJ97y3m0dA3wxf/ck+5yJAnUj17mjN/vP8kLp3r56BVnp7uUtBurj/2G\nZZXcs/EQf75mIRcvndsnqjONjuhlzrhn4yEqCnO4+rwF6S5lRrry3PlUFObwwXu3aRjjDKOglznh\neEc/v2lo5vr1tToJO47c7CDv2rCEnoEQH7hnKwOhcLpLkgRR0MuccN/mw0TcedfFOgl7JgtK8vjS\n9WvYdrhdXS4zSFxt9GZ2FfBVIAjc6e6fH/X6LcAXgeHZDG539zsTWKfIlA2GIty3+TCvPruaxZVz\nZzjiqWrvHeJVK6q5d9NhuvtDbBgxVv87X744jZXJVE0Y9GYWBL4BXAk0Ak+Z2QPuPvrf/Q/d/UNJ\nqFFkWn6+/SjNXQN86bKl6S5l1viz1fNp7urnFzuOUZKXxaqFurhsNoun6eZiYJ+7H3D3QeAHwHXJ\nLUskMSIR547HD3BuTQmvXFGV7nJmjYAZN160mNryfH7w1BEOndLAZ7NZPEG/CDgy4nFj7LnR3mpm\nO83sfjOrG2tFZnabmW0xsy0tLS1TKFdkch57vpm9zd3c9qqlc2pykUTIyQpw0yX1lBVkc/cfD6kn\nziyWqH70vwDuc/cBM3s/cBfw2tELufsdwB0A69ev14SVknT/9rsDLCzN45rzF6a7lFlpeJKSb/9u\nP3f94QXys4OU5I89ibra72eueI7ojwIjj9Br+dNJVwDc/ZS7D8Qe3glcmJjyRKZux5F2Nh1s5T2v\nWEp2UB3Mpqq8MIebL62ndyjMXX98gf4hdbucbeL57X8KWGFmS80sB7gReGDkAmZWM+LhtUBD4koU\nmZo7Hj9AcV4WN16sI83pWliWzzsvXsyJzn7u3XyYcEQfyGeTCYPe3UPAh4BfEw3wH7n7LjP7rJld\nG1vsI2a2y8x2AB8BbklWwSLxaGjq5KFnm3j3hiUU5Wqkj0Q4e34xb15Xy77mbn62/SjuCvvZIq6/\nAHd/CHho1HOfHHH/48DHE1uayNS4O//8UAOl+dm8/1VnpbucjHLhknJaewZ5dE8zNaV5XHqWejLN\nBmq4lIzz2PMtPLH3JB957QpKC8Y+cShT97pz57GqpoSHnmliX3N3usuROCjoJaOEwhH++ZcN1FcW\naMz5JAmYcf2FtVQV5XLf5sOc6h6Y+E2SVgp6ySg/2tLI3uZu/v7qleRk6dc7WXKzg9x0ST0A92w6\nxGAokt6C5Ix0lkoyRlvPIF9+eA8X1Zfz+tXRoYjHGnddEqOiMIcbLqrjrj+8wM+3H+XmS5foorQZ\nSoc8khHcnb/7yU46+ob49LWrFTgpcvb8Yl6zch5PH2nnB08dmfgNkhYKeskI92w6zMO7T/A/rlrJ\nag3AlVKvXTmPFfOK+NQDu3j2aEe6y5ExKOhl1ttzvIv//eBuLj+nmvdohMqUC5jx9vV1VBbm8IHv\nb6WjVxOMzzQKepnVugdCfPi+bRTnZfOl69cQCKjJJh0Kc7P4xrsu4HhHP3/9o+1EdOXsjKKgl1mr\ndzDErf9vMwdaevjKDWuoKspNd0lz2gWLy/mfb1zFI881863f7U93OTKCgl5mpb7BMO/93ha2Hmrj\nqzeu45UrqtNdkgA3XbKE69Yu5F//aw9P7j2Z7nIkRkEvs07vYIjb/n0LGw+e4stvX8sbz6+Z+E2S\nEmbG595yHsvnFfGh+7Zx8KQmLJkJ1I9eZoXh/vDDoyee7BrgzesW0TsYVl/5GaYgJ4v/e9N63vzN\nP/Ce7z3FTz9wKeWFOekua07TEb3MGtsOt/HNx/bROxjm1suWsr6+It0lyTiWVBZyx7sv5GhbH++/\nZysDIY1hn04Kepnxdja2850nD3D/1kZqywv48GuWs3xeUbrLkgmsr6/gi9efz+aDrfzd/Ts1hn0a\nqelGZqRwxNnyQit3//EQv3ymiYKcIG88r4YNyyoJqgvlrHHd2kU0tvXxxV/vwR3+9e1rNNtXGijo\n57AztW1PZ/7PqbSZhyPOhmUV7DnRxZYX2njomSaauwYoyAnykdcup6wgh7zs4JRrkuQb7+deXpDD\n377+HL746z0MhMJ8/R0XpHXAuWT93s9kCnpJmoFQmFPdg3T0DdHZP0RXf4jewTADQ2EGQhH6h8L0\nDYXpHQzT3R8iHJuxKDcrwGvOmccbz6/hdefOoyAnSydcZ7kPvmY5+dlBPvvgbt539xa+esNanaBN\nIQW9TFk44pzqHuBE5wDHO/s5dKqHAyd7+OP+U5zsHqCrP/SS9+RlB8jLDpKXFSQ3K0BZfjY1pfkU\n52XxpnULOWd+CcuqC3X0noHe84qlFOYG+Z8/e5arv/oEX75hjWaoShEF/RwWjjgnOvs53tnP8Y5+\nWroG6B4I0TMY4p9+uZtQxBk+fRYwyAoECBiEIs5gKEJojJNrZQXZlORls2JeEVVFuVQW5VJekE1x\nXjZFuVlnbF9/87raJH2nMlPccNFiVi8s5SP3Pc277tzEba9axgdfs5ySPM0ElkxxBb2ZXQV8FQgC\nd7r750e9ngvcDVwInAJucPcXEltq6mVaW95QOMLOxg42HjjFpoOtbDxw6vSEEVkBo7o4l+K8LOYV\n51KQEyQQMIZj2R0i7kQcggFjTV0pOcEgFUU5zC/OZV5JHosrCqgozFEzi7zE6N+Jmy6p58Gdx/i3\n3x3gvk2Hec8rlnLrZUspzU9u4IcjTlf/EJ39IfqHwoQjTiTimBn5OUH2t3RTVZibcVNQThj0ZhYE\nvgFcCTQCT5nZA+6+e8Ri7wXa3H25md0IfAG4IRkFJ1MoHKF3KEzvQJihcIS2nkEwCJqRmx0gJxiY\nVeOc9wyE2NnYwbbDbWx5oZXNB1vpGYz2Zz5nfjHr6sqoryykpiyPysLcSfVmmY3/6GTmyMkK8JYL\nann50kqeb+7i//xmL3c8foDLz6nmylXzec058ygrmFobvrvT0jXAvuZu9rV0c6Clh/0t3Rw82UNT\nR/8Zu3l+OzZGT0leFksqC1lSWcDyeUWsmFfM8nlFLKksmJXNivEc0V8M7HP3AwBm9gPgOmBk0F8H\nfDp2/37gdjMzd09Lx9mhcPREX/dAiO7+EJ39ITr6BmntGaKtZ5CTPQOc7BrkZPcArT2DtPYM0tY7\nSO/gmS/qMCA3O0BhThb3bz1CRWEuVUU5VBVFv5YX5lBWkEN5QbSZoigvi6LcLHKzggnvEhiOePT7\nGwjR1jNIc1c/zZ0DHGnrZV9zN3ubuzl0qvf0L/VZ1YW8+YJFXHpWFRuWVerIW2aEReX5/O1V57Dr\nWAf3bDzMbxpO8NAzxzGDuvICzp5fxPJ5xVQW5lCan01JfjSywhEIRSJ09kd//1t7Bjna3seR1l4a\n2/roHvjT+aHCnCBLqwu5YHE5dRX5HGntoyQvi7ycIFlmBAJGxKPjJ61bXEZzVz9HWvs43NrLzsYO\nfvlME8NJZga15fksrSpiUVkzODZIAAAIhElEQVQ+teX51JTmUVmUS0VBDmWxv/2C3OCMOjCMJ+gX\nASOnjmkEXj7eMu4eMrMOoBJI+KhGjz7XzP/6+bOnmxLCkehtKBxtMx4IRSa8MCMnGKCqKIfKolwq\ni3JYMa+I8sIcSvKyKcwNUpCTRXbQ2HjgFBBtkx4YipzuKdIzGCI/J0hjWy/bj7TT2jPARNeCZAeN\nnGCArGCA7KARDBgBizaNmBnDvw/DX92Hb9HmknDsex0MRRgMR8adozMYsOhRSHUR15xXw7rF5ayt\nK1MPB5nRVi8s5XNvOY9/iryMHY3tPLH3JHtOdLH3RBeP7WkZ83zQSMW5WdSU5VFXXsCGZZUsrSrk\nrOoils8rYn5J7osC90wHOG9at+glz/UPhdnf0s2+5uingwMne3jhZA/PHu2gtWdw3HUFA9G/+eyg\nkZMVIGB/+rt/14bF/LfLl8exZxIjpSdjzew24LbYw24z25PK7Y+0N/mbqCIJ/+jicQB4NMnbeNfU\n3nbGfTLFdWaCtP2uzATj/NwnvU+eTV4tCfcH4IOTf9vwPlky2TfGE/RHgboRj2tjz421TKOZZQGl\nRE/Kvoi73wHcMdkiZyMz2+Lu69Ndx0yifTI27ZeX0j55qensk3guT3sKWGFmS80sB7gReGDUMg8A\nN8fuvw34bbra50VE5MUmPKKPtbl/CPg10e6V33X3XWb2WWCLuz8AfAf4dzPbB7QS/WcgIiIzQFxt\n9O7+EPDQqOc+OeJ+P3B9Ykub9eZEE9UkaZ+MTfvlpbRPXmrK+8TUwiIiktk0XqiISIZT0CeImV1v\nZrvMLGJm454ZN7OrzGyPme0zs79PZY2pZmYVZvawme2NfS0fZ7mwmW2P3Uaf6M8IE/3czSzXzH4Y\ne32TmdWnvsrUi2O/3GJmLSN+P96XjjpTxcy+a2bNZjZmb1GL+lpsf+00swviWa+CPnGeBd4CPD7e\nAiOGk7gaWAW8w8xWpaa8tPh74BF3XwE8Ens8lj53Xxu7XZu68lIjzp/76WFEgK8QHUYko03i7+GH\nI34/7kxpkan3PeCqM7x+NbAidrsN+FY8K1XQJ4i7N7j7RBeAnR5Owt0HgeHhJDLVdcBdsft3AW9K\nYy3pFM/PfeS+uh94nc2U6+eTZ679PUzI3R8n2nNxPNcBd3vURqDMzGomWq+CPrXGGk7ipddcZ475\n7t4Uu38cmD/OcnlmtsXMNppZJv4ziOfn/qJhRIDhYUQyWbx/D2+NNVPcb2Z1Y7w+l0wpQzQe/SSY\n2W+ABWO89Al3/3mq65kJzrRPRj5wdzez8bp4LXH3o2a2DPitmT3j7vsTXavMSr8A7nP3ATN7P9FP\nPa9Nc02zjoJ+Etz9immuIp7hJGaVM+0TMzthZjXu3hT7eNk8zjqOxr4eMLPHgHVAJgV9woYRyTAT\n7hd3H7kP7gT+JQV1zWRTyhA13aRWPMNJZJKRQ2PcDLzkU4+ZlccmrsHMqoDLePEQ2JlAw4iMbcL9\nMqr9+VqgIYX1zUQPADfFet9sADpGNI+Oz911S8ANeDPR9rIB4ATw69jzC4GHRiz3BuB5okesn0h3\n3UneJ5VEe9vsBX4DVMSeX090pjKAS4FngB2xr+9Nd91J2hcv+bkDnwWujd3PA34M7AM2A8vSXfMM\n2S+fA3bFfj8eBVamu+Yk74/7gCZgKJYn7wX+EvjL2OtGtKfS/tjfy/p41qsrY0VEMpyabkREMpyC\nXkQkwynoRUQynIJeRCTDKehFRDKcgl5EJMMp6CUtzOzTZvYxM/usmV0Re+6VsaGet5tZvpl9Mfb4\ni0mqoX684WBHLHNfbJyV/56A7d1iZrdPdz2xddWb2TsTsS7JfBoCQdLKR0xJCbwL+Jy73wNgZrcR\nvcgqPNF6zCzLo4OBJYyZLQAu8ujQwTNNPfBO4N401yGzgI7oJWXM7BNm9ryZPQmcE3vue2b2ttiE\nEm8H/tHMvh+bgKQI2GpmN4yzvu+Z2bfNbBPwL2ZWGJu4YbOZPW1m18WWqzezJ8xsW+x2aZwl/xew\nKPYJ45Vm9pjFJpUxsyozeyF2/xYz+6mZ/adFJ1k5PR6Lmd0a+543Ex3e4Uz7p9rMfmJmT8Vul8We\nf/WIiTeeNrNi4PPAK2PPTfvThmQ2HdFLSpjZhUTHMllL9PduG7B1+HV3v9PMXgE86O73x97T7e5r\nJ1h1LXCpu4fN7J+JjhHzHjMrAzbHRtdsBq50934zW0H0MvNxZwEb4dpYPWtj9Zxp2bVEB2MbAPaY\n2deBEPAZ4EKiww4/Cjx9hnV8FfiKuz9pZouBXwPnAh8DPujuvzezIqCf6CQuH3P3a+L4PmSOU9BL\nqrwS+A937wWwxE0Z+OMRTTt/BlxrZh+LPc4DFgPHgNvNbC0QBs5O0LZHesTdOwDMbDewBKgCHnP3\nltjzP5xg21cAq0b8QymJBfvvgS+b2feBn7p7Y+bPSSKJpKCX2a5nxH0D3uqjZvoys08THWhuDdHm\nyv4pbivEn5o780a9NjDifpip/W0FgA3uPrq+z5vZL4kOAPZ7M3v9FNYtc5ja6CVVHgfeFOtNUwz8\neRK28WvgwxY73DWzdbHnS4Emd48A7waCU1z/C0SbYSA6lPBENgGvNrNKM8sGrp9g+f8CPjz8IPYJ\nBDM7y92fcfcvEB3adyXQBRRPrnyZqxT0khLuvg34IdHhZn9FNLAS7R+BbGCnme2KPQb4JnCzme0g\nGpI947x/Il8CPmBmTxNtljkjj44T/mngj0SbXyYaS/0jwPpYd87dRIenBfiomT1rZjuJDl/7K2An\nEDazHToZKxPRMMUiIhlOR/QiIhlOJ2NlxjOzT/DS9u0fu/s/JWj9rwe+MOrpg+7+5kSsf4ztJfX7\nERlNTTciIhlOTTciIhlOQS8ikuEU9CIiGU5BLyKS4RT0IiIZ7v8DKY/7MjGVzFcAAAAASUVORK5C\nYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11a191a90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "% matplotlib inline\n",
    "sns.distplot(nfyydf[\"diff_real_fund_est\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.05"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len([i for i in nfyydf[\"diff_real_fund_est\"][-100:] if abs(i)>0.5])/100"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "可以看出有大约5%的概率，预测净值和真实净值的偏差超过0.5个百分点，总体上是一个可以接受的净值预测\n",
    "可以和其他qdii基金比一下预测精度"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "271 12 0.04240282685512368\n"
     ]
    }
   ],
   "source": [
    "# 反向预测的比例\n",
    "p = 0\n",
    "n = 0\n",
    "for i, row in nfyydf.iterrows():\n",
    "    r = row[\"fund_est\"] * row[\"real\"]\n",
    "    if r >= 0:\n",
    "        p += 1\n",
    "    else:\n",
    "        n += 1\n",
    "print(p, n, n / (n + p))\n",
    "# 竟然有4%+的机会预测完全反号。。。。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 国泰商品净值预估"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "gtspdf = estimate_table(\n",
    "    \"20190101\",\n",
    "    \"20200305\",\n",
    "    (\"real\", \"\", {\"F160216\": 100}),\n",
    "    (\"fund_est\", \"USD/CNY\", holdings_160216_19s4),\n",
    "    (\"fund_est_2111\", \"2111\", holdings_160216_19s4),\n",
    "    _default=cache_table,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>date</th>\n",
       "      <th>real</th>\n",
       "      <th>fund_est</th>\n",
       "      <th>fund_est_2111</th>\n",
       "      <th>diff_real_fund_est</th>\n",
       "      <th>diff_real_fund_est_2111</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>263</th>\n",
       "      <td>2020-02-07</td>\n",
       "      <td>-1.173709</td>\n",
       "      <td>-1.508188</td>\n",
       "      <td>-0.765309</td>\n",
       "      <td>0.334479</td>\n",
       "      <td>-0.408400</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>264</th>\n",
       "      <td>2020-02-10</td>\n",
       "      <td>-1.425178</td>\n",
       "      <td>-1.450684</td>\n",
       "      <td>-1.825055</td>\n",
       "      <td>0.025506</td>\n",
       "      <td>0.399877</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>265</th>\n",
       "      <td>2020-02-11</td>\n",
       "      <td>0.963855</td>\n",
       "      <td>0.941957</td>\n",
       "      <td>0.627059</td>\n",
       "      <td>0.021899</td>\n",
       "      <td>0.336796</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>266</th>\n",
       "      <td>2020-02-12</td>\n",
       "      <td>2.147971</td>\n",
       "      <td>2.610018</td>\n",
       "      <td>2.960598</td>\n",
       "      <td>-0.462046</td>\n",
       "      <td>-0.812627</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>267</th>\n",
       "      <td>2020-02-13</td>\n",
       "      <td>0.233645</td>\n",
       "      <td>0.378472</td>\n",
       "      <td>0.355457</td>\n",
       "      <td>-0.144827</td>\n",
       "      <td>-0.121812</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>268</th>\n",
       "      <td>2020-02-14</td>\n",
       "      <td>1.165501</td>\n",
       "      <td>1.056623</td>\n",
       "      <td>1.115977</td>\n",
       "      <td>0.108878</td>\n",
       "      <td>0.049524</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>269</th>\n",
       "      <td>2020-02-17</td>\n",
       "      <td>-0.230415</td>\n",
       "      <td>-0.068726</td>\n",
       "      <td>-0.080150</td>\n",
       "      <td>-0.161689</td>\n",
       "      <td>-0.150265</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>270</th>\n",
       "      <td>2020-02-18</td>\n",
       "      <td>0.230947</td>\n",
       "      <td>0.149082</td>\n",
       "      <td>0.332609</td>\n",
       "      <td>0.081865</td>\n",
       "      <td>-0.101663</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>271</th>\n",
       "      <td>2020-02-19</td>\n",
       "      <td>2.304147</td>\n",
       "      <td>2.297944</td>\n",
       "      <td>2.040751</td>\n",
       "      <td>0.006204</td>\n",
       "      <td>0.263396</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>272</th>\n",
       "      <td>2020-02-20</td>\n",
       "      <td>0.675676</td>\n",
       "      <td>0.739045</td>\n",
       "      <td>1.074390</td>\n",
       "      <td>-0.063369</td>\n",
       "      <td>-0.398714</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>273</th>\n",
       "      <td>2020-02-21</td>\n",
       "      <td>-0.894855</td>\n",
       "      <td>-0.966361</td>\n",
       "      <td>-1.166828</td>\n",
       "      <td>0.071506</td>\n",
       "      <td>0.271973</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>274</th>\n",
       "      <td>2020-02-24</td>\n",
       "      <td>-3.611738</td>\n",
       "      <td>-3.480442</td>\n",
       "      <td>-3.498332</td>\n",
       "      <td>-0.131296</td>\n",
       "      <td>-0.113406</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>275</th>\n",
       "      <td>2020-02-25</td>\n",
       "      <td>-3.044496</td>\n",
       "      <td>-3.001084</td>\n",
       "      <td>-3.201197</td>\n",
       "      <td>-0.043412</td>\n",
       "      <td>0.156700</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>276</th>\n",
       "      <td>2020-02-26</td>\n",
       "      <td>-1.932367</td>\n",
       "      <td>-2.232813</td>\n",
       "      <td>-1.959383</td>\n",
       "      <td>0.300446</td>\n",
       "      <td>0.027016</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>277</th>\n",
       "      <td>2020-02-27</td>\n",
       "      <td>-3.448276</td>\n",
       "      <td>-3.543438</td>\n",
       "      <td>-3.912623</td>\n",
       "      <td>0.095162</td>\n",
       "      <td>0.464347</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>278</th>\n",
       "      <td>2020-02-28</td>\n",
       "      <td>-3.061224</td>\n",
       "      <td>-3.473936</td>\n",
       "      <td>-3.441291</td>\n",
       "      <td>0.412712</td>\n",
       "      <td>0.380066</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>279</th>\n",
       "      <td>2020-03-02</td>\n",
       "      <td>3.684211</td>\n",
       "      <td>4.259999</td>\n",
       "      <td>4.176892</td>\n",
       "      <td>-0.575789</td>\n",
       "      <td>-0.492682</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>280</th>\n",
       "      <td>2020-03-03</td>\n",
       "      <td>-0.761421</td>\n",
       "      <td>-0.858441</td>\n",
       "      <td>-0.423418</td>\n",
       "      <td>0.097020</td>\n",
       "      <td>-0.338003</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>281</th>\n",
       "      <td>2020-03-04</td>\n",
       "      <td>-0.255754</td>\n",
       "      <td>-0.283699</td>\n",
       "      <td>-0.793606</td>\n",
       "      <td>0.027944</td>\n",
       "      <td>0.537851</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>282</th>\n",
       "      <td>2020-03-05</td>\n",
       "      <td>-2.564103</td>\n",
       "      <td>-2.663462</td>\n",
       "      <td>-2.351544</td>\n",
       "      <td>0.099360</td>\n",
       "      <td>-0.212558</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          date      real  fund_est  fund_est_2111  diff_real_fund_est  \\\n",
       "263 2020-02-07 -1.173709 -1.508188      -0.765309            0.334479   \n",
       "264 2020-02-10 -1.425178 -1.450684      -1.825055            0.025506   \n",
       "265 2020-02-11  0.963855  0.941957       0.627059            0.021899   \n",
       "266 2020-02-12  2.147971  2.610018       2.960598           -0.462046   \n",
       "267 2020-02-13  0.233645  0.378472       0.355457           -0.144827   \n",
       "268 2020-02-14  1.165501  1.056623       1.115977            0.108878   \n",
       "269 2020-02-17 -0.230415 -0.068726      -0.080150           -0.161689   \n",
       "270 2020-02-18  0.230947  0.149082       0.332609            0.081865   \n",
       "271 2020-02-19  2.304147  2.297944       2.040751            0.006204   \n",
       "272 2020-02-20  0.675676  0.739045       1.074390           -0.063369   \n",
       "273 2020-02-21 -0.894855 -0.966361      -1.166828            0.071506   \n",
       "274 2020-02-24 -3.611738 -3.480442      -3.498332           -0.131296   \n",
       "275 2020-02-25 -3.044496 -3.001084      -3.201197           -0.043412   \n",
       "276 2020-02-26 -1.932367 -2.232813      -1.959383            0.300446   \n",
       "277 2020-02-27 -3.448276 -3.543438      -3.912623            0.095162   \n",
       "278 2020-02-28 -3.061224 -3.473936      -3.441291            0.412712   \n",
       "279 2020-03-02  3.684211  4.259999       4.176892           -0.575789   \n",
       "280 2020-03-03 -0.761421 -0.858441      -0.423418            0.097020   \n",
       "281 2020-03-04 -0.255754 -0.283699      -0.793606            0.027944   \n",
       "282 2020-03-05 -2.564103 -2.663462      -2.351544            0.099360   \n",
       "\n",
       "     diff_real_fund_est_2111  \n",
       "263                -0.408400  \n",
       "264                 0.399877  \n",
       "265                 0.336796  \n",
       "266                -0.812627  \n",
       "267                -0.121812  \n",
       "268                 0.049524  \n",
       "269                -0.150265  \n",
       "270                -0.101663  \n",
       "271                 0.263396  \n",
       "272                -0.398714  \n",
       "273                 0.271973  \n",
       "274                -0.113406  \n",
       "275                 0.156700  \n",
       "276                 0.027016  \n",
       "277                 0.464347  \n",
       "278                 0.380066  \n",
       "279                -0.492682  \n",
       "280                -0.338003  \n",
       "281                 0.537851  \n",
       "282                -0.212558  "
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gtspdf.iloc[-20:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.10589187603833962, 0.22399439609501928)"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(\n",
    "    np.mean(np.abs(gtspdf[\"diff_real_fund_est\"][-100:])),\n",
    "    np.mean(np.abs(gtspdf[\"diff_real_fund_est_2111\"][-100:])),\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(-0.02266683846031749, -0.023893072096493828, 0.007916101144943788)"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 看看有没有肉眼可见的系统偏差\n",
    "(\n",
    "    np.mean(gtspdf[\"diff_real_fund_est\"][-150:-100]),\n",
    "    np.mean(gtspdf[\"diff_real_fund_est\"][-100:-50]),\n",
    "    np.mean(gtspdf[\"diff_real_fund_est\"][-50:]),\n",
    ")\n",
    "# 最近的最准，也可以理解，毕竟是按照4季度底持仓模拟的"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "预测精度要比南方原油好，用人民币中间价，平均误差只有0.1个百分点。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x1131cf4a8>"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAELCAYAAAAiIMZEAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4yLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvNQv5yAAAIABJREFUeJzt3XmY3FWd7/H3t5be9z3pNfsCgQTC\nDoI4uAveR1QG1xkU13F0xnuvM14dR2fmjo7XGRVHL6NeVxQX1IwggiCDgISE7J0FOp30nk7v+1ZV\n3/tHVYem6e6q7lTVr5bv63nqSS0nv/qkn6pvTp/z+50jqooxxpjU4nI6gDHGmOiz4m6MMSnIirsx\nxqQgK+7GGJOCrLgbY0wKsuJujDEpyIq7McakICvuxhiTgqy4G2NMCvI49cZlZWXa0NDg1NsbY0xS\nevbZZ3tVtTxcO8eKe0NDA3v37nXq7Y0xJimJSEsk7WxYxhhjUpAVd2OMSUFW3I0xJgVZcTfGmBRk\nxd0YY1KQFXdjjElBVtyNMSYFWXE3xpgUZMXdGGNSkGNXqBoTT/fsbo3ZsW+/oi5mxzZmpaznbowx\nKciKuzHGpCAr7sYYk4KsuBtjTAqy4m6MMSnIzpYxZp4Zf4Anm3rpHJygsjCL1YXZ1JbkkJdpXxeT\nPOzTakyIqnKkc5jfHOlicHyGomwvjZ3DKOB1C2/dWcfW1QVOxzQmIlbcjQn5+b4O9rUOUFWQxXuu\nrWFteR5TPj9nhia5/3AXP9zdwqsvrOLa9WWIiNNxjVmSFXdjgANtg+xrHeBlG8p45QVVuELFO9Pj\npr40l/det5afPtvOb46coXd0ilu2V59rY0wisglVk/YGx6fZdbCDupIcbtpatWDR9rpd3HZZLddv\nLGfP6QH+eLLPgaTGRC5scReRLBF5RkQOikijiPz9Am0yReReEWkSkd0i0hCLsMZEW0CVnz3bTkDh\nLTtrcbsW7427RHjl1kq2VOXz4JEzdAxOxDGpMcsTSc99CrhRVS8GtgOvFpEr57W5AxhQ1fXAvwKf\nj25MY2LjqaZemnvHeMNFqyjJzQjbXkR40yU15Ga6uXdPK1M+fxxSGrN8YYu7Bo2GHnpDN53X7Bbg\nu6H7PwNeITbjZBLctC/AoyfOsqkyn0vqiiP+ezmZHt6ys5a+0Wn+82BXDBMas3IRjbmLiFtEDgBn\ngYdVdfe8JtVAG4Cq+oAhoDSaQY2Jtv1tA0zOBLh+Y/myz35ZW57H9ZvK2dc6wBPP98YooTErF1Fx\nV1W/qm4HaoDLReTClbyZiNwpIntFZG9PT89KDmFMVKgqT53sY3VRFvWlOSs6xss3VVCc4+Uf7j+K\nPzD/l1ljnLWss2VUdRD4PfDqeS91ALUAIuIBCoGXnE6gqner6k5V3VleXr6yxMZEQdPZUXpGprh6\n3crPWfe6XbzqgiqOnxnh58+2RzmhMecnkrNlykWkKHQ/G7gJOD6v2S7gXaH7twKPqqp1ZUzCeupk\nH3mZHi6qLjyv42yrLmRHXRFffOgEY1O+KKUz5vxF0nNfBfxeRA4BewiOuf9aRD4rIjeH2nwLKBWR\nJuCvgE/EJq4x5693ZIoT3SNcvqYEj/v8LvUQEf7X67ZwdmSKux9vjlJCY85f2CtUVfUQsGOB5z89\n5/4k8OboRjMmNp5q7sPtEq5YUxKV411aX8Lrtq3i7sebedsVdVQUZEXluMacD7tC1aQVf0A50DbA\nhasLyM/yRu24//1Vm5jy+fnWE6eidkxjzocVd5NWmntHmZwJcFFNUVSP21CWyxsuXs0Pnm5hcHw6\nqsc2ZiWsuJu00tg5TIbbxfqKvKgf+wM3rGNs2s93n2qJ+rGNWS5bFdKkjYAqRzuH2ViVj/c8J1Ln\numd367n7W6ry+cZ/naQg20Omx/2StrdfURe19zVmKdZzN2mjrX+c0SkfF8Rww43rN1UwMeNnz6n+\nmL2HMZGw4m7SxpGOIdwuYVNlfszeo64kh7VluTzR1IvPH4jZ+xgTjhV3kxZUlcauYTZU5JHlfelw\nSTRdv6mc4UkfB9uHYvo+xizFirtJC51DkwyOz8R0SGbW+vI8KvIz+ePJXuxCbeMUK+4mLTR2DuES\n2FwV++IuIly1rpTOoUla+8dj/n7GLMSKu0kLRzuHaSjLJTczPieI7agtJsvr4inbjs84xIq7SXld\nQxOcHZlicwwnUufL8LjYWV9CY+cQwxMzcXtfY2ZZcTcp78mmYO95XQwuXFrKlWtLUYXddlqkcYAV\nd5PynmrqJTfDTWWcF/Qqyc1gU1U+z5zut9MiTdxZcTcpTVV5oqmXdRV5uBzY1veqtaWMTfk40jkc\n9/c26c2Ku0lpJ3tGOTsyxbry+A7JzFpXkUdxjpe9LTY0Y+LLirtJaefG2x0q7i4RLq0vprlnjP4x\nWy3SxI8Vd5PSnmjqpa4kh5LcDMcyXFJXjAD7Wgccy2DSjxV3k7J8/gBPN/dxzfpSR3MU5WSwviKP\nZ1sG8AfsilUTH1bcTco60jnMyKSPa9aXOR2FS+uLGZqY4cmmXqejmDRhxd2krNlCetVaZ3vuAFtX\nFZDtdfOTvW1ORzFpwoq7SVlPNvWyZVUBpXmZTkfB43axva6Ihxq7bRs+Exe2E5NJSVM+P3tbBnjn\nlfVORzlnZ30xfzzZx6d+1RjRbxO2a5M5H9ZzNynpcPsQ074Al68pcTrKOasKs6nIz+SwrfNu4iBs\ncReRWhH5vYgcFZFGEfnLBdrcICJDInIgdPt0bOIaE5m9LcHTDi+tL3Y4yYttqymkpW/MFhMzMRdJ\nz90H/LWqbgWuBD4kIlsXaPcHVd0eun02qimNWaa9p/tZW56bEOPtc22rLkSBI53WezexFba4q2qX\nqu4L3R8BjgHVsQ5mzEoFAsrelgEuq0+cIZlZFflZVBVkcciGZkyMLWtCVUQagB3A7gVevkpEDgKd\nwMdVtfG80xmzhHt2ty74/Nnh4JZ6vkBg0TZO2lZTyMNHg2fNFOU4d+WsSW0RT6iKSB7wc+Cjqjp/\nibt9QL2qXgx8FfjlIse4U0T2isjenp6elWY2ZkktfcGt7epLcx1OsrBt1YUAtlKkiamIiruIeAkW\n9h+q6n3zX1fVYVUdDd1/APCKyEsuC1TVu1V1p6ruLC8vP8/oxiyspX+M3Aw3pQ6uJ7OUsrxMVhdm\ncbh90OkoJoVFcraMAN8CjqnqlxZpUxVqh4hcHjqubR5pHHG6b5z60lzEgfXbI7WtupC2gQkGbKVI\nEyOR9NyvAd4B3DjnVMfXisj7ReT9oTa3AkdCY+5fAW5TVVshycTdyOQM/WPT1JfmOB1lSdtqigA7\na8bETtgJVVV9AliyC6SqdwF3RSuUMSs1O97ekKDj7bNKcjOoKsjixJkRrttgQ5Qm+uwKVZNSWvrG\n8LiEVUXx3S91JTZW5nO6b4zJGb/TUUwKsuJuUkpL/zi1JTl4XIn/0d5UlU9AoensqNNRTApK/G+A\nMRGa9gXoHJygviSxx9tn1ZXkkOV18Vz3iNNRTAqy4m5SRsfgBAENFs1k4HYJ6yvyOdE9gp1/YKLN\nirtJGW39wcnUmiQp7gCbKvMZmfTRNTTpdBSTYqy4m5TRNjBOSW4GeZnJs03Bxso8ABuaMVFnxd2k\njLb+cWqLs52OsSz5WV5WF2Vxwoq7iTIr7iYlDE3MMDzpozaJhmRmbarMp7VvnIlpOyXSRI8Vd5MS\nZsfba4uTs7gr8PxZ672b6LHiblJC28A4bpewqjDxL16ar6Ykh2yvm+e77Xx3Ez1W3E1KaOsfZ3Vh\nFh538n2kXSI0lOVyqm/M6SgmhSTfN8GYefwBpWNwIinH22etKculf2yaIdtb1USJFXeT9LqHJ5nx\na1KOt89aUxZc6OxUrw3NmOiw4m6SXttAaDI1iXvuqwqzyPK6ONVrQzMmOqy4m6TX1j9Oboab4hyv\n01FWzCVCfUmuFXcTNVbcTdJr6w+OtyfyzkuRWFOWS+/oNMOTNu5uzp8Vd5PUJqb99IxOJfWQzKzZ\ncffT1ns3UWDF3SS19sHQYmFJtuzAQlYXZZPhsXF3Ex1W3E1Sax+YAKCmKPl77m6XUF+SQ7MVdxMF\nVtxNUmsfmKAsL4PsDLfTUaJiTVkuPSNTjE75nI5ikpwVd5O0VJX2/nFqkvj89vleON/deu/m/Fhx\nN0lreNLHyJQvJcbbZ1UXZ+N1ixV3c96suJuk1T4wO5maOj13j8tFTXHOuVUujVmpsMVdRGpF5Pci\nclREGkXkLxdoIyLyFRFpEpFDInJJbOIa84L2gQlcQlKuBLmU2uIcuoYmmJyx9d3NykXSc/cBf62q\nW4ErgQ+JyNZ5bV4DbAjd7gS+HtWUxiygbWCcVYXZeJNwJcil1JVkE1Bo7BxyOopJYmG/Farapar7\nQvdHgGNA9bxmtwDf06CngSIRWRX1tMaEBAJKx8AE1Sk03j5rdoPv/a2DDicxyWxZXR4RaQB2ALvn\nvVQNtM153M5L/wMwJmqae8eY8gWSbs/USBRkeSnK9rK/zYq7WbmIi7uI5AE/Bz6qqsMreTMRuVNE\n9orI3p6enpUcwhgADoYKXypNps5VU5LDAeu5m/MQUXEXES/Bwv5DVb1vgSYdQO2cxzWh515EVe9W\n1Z2qurO8vHwleY0B4FD7IBkeF+X5mU5HiYm64mw6Bic4OzLpdBSTpCI5W0aAbwHHVPVLizTbBbwz\ndNbMlcCQqnZFMacxL3KgfYjqomxcSb4S5GJmF0Kz3rtZqUh67tcA7wBuFJEDodtrReT9IvL+UJsH\ngGagCfgP4IOxiWsMTPn8HOscTqmLl+ZbXZSNxyUcsHF3s0KecA1U9Qlgye6RqirwoWiFMmYpx7tG\nmPYHUna8HcDrdrFlVYGdMWNWLLVOEDZp4VB7sOCl4pkyc22vLeJQ+yD+gDodxSQhK+4m6RxoG6Is\nL4PC7OTdVi8SO+qKGJv203TWNs02y2fF3SSdg+2DXFxTlPTb6oWzvbYIgP2tAw4nMcnIirtJKsOT\nM5zsGeXiUOFLZWvKcinM9tq4u1kRK+4mqRxpH0KVtCjuIsLFtUUc6rA1ZszyWXE3SeVge7DQXVRd\n6HCS+NhWXcDz3SO2QqRZNivuJqkcbBukvjSH4twMp6PExbbqQnwB5fiZEaejmCRjxd0kldnJ1HRx\nYeg3lMM2NGOWyYq7SRrdw5N0DU2mxXj7rOqibIpzvBxut0lVszxW3E3SmF0Jcntteoy3Q3BS9cLq\nQg53rGghVpPGrLibpHGofQi3S9i6Kn2KOwTH3W1S1SyXFXeTNA62D7KpMp/sDLfTUeLKJlXNSlhx\nN0khEFAOtg2m1Xj7LJtUNSthxd0khdN9YwxP+tJqvH1WTXE2RTlejrRbcTeRs+JuksLB0NkiF6XR\naZCzRIRt1YXWczfLYsXdJIX9rYPkZrjZWJnvdBRHbKsu5DmbVDXLYMXdJIX9rYNcVFOE25XaK0Eu\nxiZVzXJZcTcJb3LGz7GuYXbUpd+QzCybVDXLZcXdJLwjHUP4AsqOumKnozjGJlXNcllxNwlvdj3z\n7Wl4GuQsm1Q1y2XF3SS8/W0D1BRnU56f6XQUR11YXcjzZ0eY8tmkqgnPirtJeAdaB9N6SGbWhasL\nmfErz52xPVVNeFbcTUI7MzRJ59BkWg/JzLqwugCAI502NGPCC1vcReTbInJWRI4s8voNIjIkIgdC\nt09HP6ZJVwfagptDp/OZMrPqSnLIz/JwxMbdTQQ8EbT5DnAX8L0l2vxBVV8flUTGzLG/dZAMt4sL\nVhc4HcVxIsIFqws40mnL/5rwwvbcVfVxoD8OWYx5if1tg2xZXUCmJ71WglzMhasLOdY1zIw/4HQU\nk+Ai6blH4ioROQh0Ah9X1cYoHdekmXt2t5677w8o+1sH2NlQ8qLn09mF1YVM+wKc7Bllc5X9NmMW\nF40J1X1AvapeDHwV+OViDUXkThHZKyJ7e3p6ovDWJpV1D08y41fqinOcjpIwzk2q2s5MJozzLu6q\nOqyqo6H7DwBeESlbpO3dqrpTVXeWl5ef71ubFNfaPw5AbYkV91lryvLI9rptUtWEdd7FXUSqRERC\n9y8PHbPvfI9rTEvfGPlZHopzvE5HSRhul7B1dQGNdjqkCSPsmLuI/Ai4ASgTkXbg7wAvgKp+A7gV\n+ICI+IAJ4DZV1ZglNmmjpW+c+tJcQn0HE7KtupCf7G0jEFBcabpKpgkvbHFX1T8N8/pdBE+VNCZq\nBsenGZyY4dpSG5KZ74LVBYxP+2nuHWN9RZ7TcUyCsitUTUJq6QuOt9eX5jqcJPHMLv9rQzNmKVbc\nTUJq6R8jw+2iqiDL6SgJZ31FHhkel02qmiVZcTcJqaVvnLqSnLTdeWkpXreLLVX5tvyvWZIVd5Nw\nJmf8nBmapN7G2xd1QXUhjZ3DBAJ27oJZmBV3k3Ba+8dRbLx9KRdVFzIy6aMldC2AMfNZcTcJp6Vv\nDJdAbUm201ES1raa4KTqofZBh5OYRGXF3SSc033jrCrMtsXClrCxMp8Mj4vDtqeqWYQVd5NQ/AGl\nfWDcxtvD8LpdbF1VwCGbVDWLiNaqkMZERefgBDN+tfF2CLsSZpbXxb7WQX7wdAtvv7I+TqlMsrCe\nu0kop/vGAKznHoHqohymfQF6R6acjmISkBV3k1Cae8Yoy8ukIMsWCwunpjg44dwxOOFwEpOIrLib\nhDHjD3Cqb4x15TYkE4ny/Ey8bqHdirtZgBV3kzAOtQ8y7QuwrtwWw4qES4TVRdl0DFhxNy9lxd0k\njCeb+hBgbZn13CNVU5RN19AEPttT1cxjxd0kjKdO9rKqMIucTDuJK1LVxdnM+JWmnlGno5gEY8Xd\nJISJaT/7WgZZa0Myy1JdFDyr6JBdzGTmseJuEsKzLQNM+228fblK8zLItCtVzQKsuJuE8OTJXjwu\noaHMzm9fjtlJVbtS1cxnxd0khKdO9rG9tsjWk1mBmqJsjnUOM+XzOx3FJBAr7sZxw5MzHG4f5Op1\npU5HSUq1JTlM+wMc7Rx2OopJIFbcjeN2N/cTULh6fZnTUZJSXUlwKGtfqy3/a15gxd047onne8jy\nuthRV+R0lKRUkO1ldWEW+1oHnI5iEogVd+MoVeV3x85y7fpyG28/DzvqizlgPXczR9jiLiLfFpGz\nInJkkddFRL4iIk0ickhELol+TJOqnusepWNwgldsqXA6SlK7pK6YjsEJuocnnY5iEkQkPffvAK9e\n4vXXABtCtzuBr59/LJMuHjneDcCNm624n4/ZIa39NjRjQsIWd1V9HOhfosktwPc06GmgSERWRSug\nSW2PHjvLtupCKguynI6S1C5YXUCG22WTquacaIy5VwNtcx63h54zZkn9Y9Psax2wXnsUZHrcXFBd\nwL4W67mboLhOqIrInSKyV0T29vT0xPOtTQJ67MRZAoqNt0fJJXXFHO4YYtpnK0Sa6BT3DqB2zuOa\n0HMvoap3q+pOVd1ZXl4ehbc2yeyR42epyM/kwtWFTkdJCZfUFTPlC3Csyy5mMtEp7ruAd4bOmrkS\nGFLVrigc16SwGX+Ax0/0cOPmClwucTpOSrBJVTNXJKdC/gj4I7BJRNpF5A4Reb+IvD/U5AGgGWgC\n/gP4YMzSmpSx51Q/I1M+G2+PotVF2VQVZNmkqgEg7K4IqvqnYV5X4ENRS2TSwsPHusnwuLh2gy05\nEE2X1BfZlaoGsCtUjQP8AeX+Q13csLGcnAzbdSmaLqkrpn1ggq4h21c13VlxN3H3zKl+zo5McfP2\n1U5HSTlXrg2urLm7ealLU0w6sOJu4m7XwU5yMty8YnOl01FSzpZVBRRkeXi6uc/pKMZhVtxNXE37\nAvzmSBc3ba0kO8MWCos2t0u4fE0pf7TinvasuJu4eqKph8HxGW6+2IZkYuWqdaW09I3TOWjj7unM\niruJq10HOinM9nLdBruILVauXFsCwO5T1ntPZ1bcTdxMTPt5+Gg3r91WRYbHPnqxsqWqgMJsL0+f\ntEnVdGbfMBM3jx4/y9i0nzfYkExMuVzCFWtKbNw9zVlxN3Hzi/0dVORncsUa2wg71q5cW0pr/zgd\nNu6etqy4m7joGprg0ePd3HppDW5bSybmXjjf3Xrv6cqKu4mLe/e0ocCfXl7ndJS0sLkqn6Icr53v\nnsasuJuY8/kD3Lunjes2lFNbkuN0nLRg4+7GiruJucdO9NA1NMntl9eGb2yi5qq1pbT1T9DaN+50\nFOMAW7XJxNwXHzpBfqaHnpFp7tnd6nSclLTQz3Vk0gfAF357nKvXvbD65u1X2NBYOrCeu4mpzsEJ\nTpwZ4dKGYptIjbPSvEzK8jI5cWbE6SjGAVbcTUz9eE9w7/TLGkocTpKeNlfl09w7xpTP73QUE2dW\n3E3MTM74uWd3Cxsr8ynOyXA6TlraXJWPP6CcPDvqdBQTZ1bcTcz8Yn8HvaPTttuSg+pLc8nyujhu\nQzNpx4q7iYlAQPmPPzRzYXUBa8tynY6TttwuYUNFPie6RwioOh3HxJEVdxMTjxw/S3PPGO+9bi0i\nNpHqpM1V+YxM+uganHQ6iokjK+4mJu5+/CTVRdm8btsqp6OkvQ2V+Qhw/Myw01FMHFlxN1G3r3WA\nPacHuOPaNXjc9hFzWl6mh9qSHBt3TzP2zTNRd/d/NVOQ5eGtl9kVqYlic1U+HYMTDE/MOB3FxElE\nxV1EXi0iJ0SkSUQ+scDr7xaRHhE5ELq9J/pRTTI41jXMg41neOdVDeRm2gXQieKC1YUAHGofdDiJ\niZew3z4RcQNfA24C2oE9IrJLVY/Oa3qvqn44BhlNEvny754nP9PDe65b43QUM0d5fiY1xdnsb7Pi\nni4i6blfDjSparOqTgM/Bm6JbSyTjBo7h3iw8Qx/du0aiuyipYSzvbaIrqFJW44gTURS3KuBtjmP\n20PPzfcmETkkIj8TERtsTUP/9rvnyc/ycMe11mtPRBfVFOESuG9/u9NRTBxEa1D0P4EfqeqUiLwP\n+C5w4/xGInIncCdAXZ2tTJfM5q9C2DEwwcNHu3nFlgruP9TlUCqzlLxMDxsq8vnV/k7+x6s220Ju\nKS6SnnsHMLcnXhN67hxV7VPVqdDDbwKXLnQgVb1bVXeq6s7y8vKV5DUJ6pHj3WR5XVyzzpYaSGQ7\n6oo4MzxpOzSlgUiK+x5gg4isEZEM4DZg19wGIjL3SpWbgWPRi2gSXXPvKMfPjPCyDeVked1OxzFL\n2LKqgPxMD/ft6wjf2CS1sMVdVX3Ah4HfEizaP1HVRhH5rIjcHGr2ERFpFJGDwEeAd8cqsEksAVV+\nc/gMhdlerllvvfZE53W7eM22Kh480sXEtC0DnMoiOs9dVR9Q1Y2quk5V/zH03KdVdVfo/t+o6gWq\nerGqvlxVj8cytEkcB9sG6Ric4JVbK/Ha1ahJ4dZLaxmb9tvEaoqzb6NZsRl/gIeOdlNdlM3FtUVO\nxzERuqyhmItrCrn78Wb8AVspMlVZcTcr9mRTL0MTM7zmwipctvJj0hARPnDDOlr6xvnNETuzKVVZ\ncTcrMjQxw2PP9bClKp+15XlOxzHLdNPWKtaW5fL1x06its57SrLiblbk14c6CQSU11202ukoZgXc\nLuF916+lsXOYJ5p6nY5jYsCKu1m2R49309g5zI2bKyjJtWUGktUbd1RTWZDJ1x876XQUEwNW3M2y\njE/7+NQvG6nIz7S9UZNcpsfNHdeu4amTfew93e90HBNlVtzNsnz5d8/TMTjBLdur8bjs45Psbr+i\nnqqCLD75iyNM+wJOxzFRZN9OE7H9rQN884lTvHVnLWts0+uUkJfp4XNvvJAT3SPc/bgNz6QSK+4m\nIqNTPv7yxweoKsjib1+3xek4Jopu2lrJ67at4iuPNtHcM+p0HBMlVtxNRP7uV420D4zzb7dtpzDb\n63QcE2V/d/NWsjwu/ua+wwTswqaUYMXdhPWfBzv5+b52Pvzy9VzWUOJ0HBMDFflZ/O1rt7D7VD//\n/liT03FMFNgml2ZJp3rH+OQvDrOjroiPvGKD03FMDL31slqebu7jiw89R1FOBm+/st7pSOY8WHE3\nixocn+bPv7MHj9vFV27bgccWBktpIsK/vPliRiZ9fOpXR8jP8nDL9oU2XTPJwIq7WdC0L8D7vv8s\nHQMT3PPeK6gtyXE6komS+btozfeyjeU0947xsXsPMDnj5y07axFbOyjpWFfMvISq8slfHGb3qX6+\ncOtF7LRx9rTidbt4x5X1NJTm8j9/fpj3fm8vZ0cmnY5llsmKu3mRQED59K8a+emz7XzkxvW8cYf9\nWp6Osrxu/vzaNXzq9Vv5w/O9vPJfH+ebf2hmcHza6WgmQlbczTk+f4CP//Qg33+6hfddv5aP3bTR\n6UjGQS4R7rh2Dfd/5Do2VubzD/cf4/J/eoS/uvcADx/tpn/MCn0iszF3A8DkjJ+P/vgADzae4eOv\n3MiHXr7exlkNAOsr8vjJ+67iaOcw9zzTwi/3d3Lf/uAerGvLctmyqoCa4mxqSnKozM+kKCeDohwv\n+VkeMj1usrwuPC4Xsx+ngCo+vzLjDzDtDzDtCzDjV3z+ACKCS4JDQ4U5XvIzPfY5XCFxai3nnTt3\n6t69ex15b/NiRzuH+ciP99N0dpTXX7SKq9fZgmBmcTP+AB0DE7T0j9PaN0bP6BQD4zMx2dUpw+2i\nNC+D3AwPVYVZVBVkUV+aQ0luxqJF//Yr6qKeI5GIyLOqujNcO+u5pzF/QPl/T57iCw+eoDDHy59d\n3cCGynynY5kE53W7aCjLpaEsFygHgr3x0UkfI1M+xqd9TEz7mZoJMBMI4PMr/jmdSCG4nvzla0rw\nuF1kul1keFy4XYISnPeZ9gcYHJ+mb2yanuEpdp/q5+nmPnyh/0CKcrysK89jU2U+GyrzyPS44/+D\nSHBW3NOQP6D8+lAnX320iaazo9y0tZLPv+kiHjxyxuloJkm5RCjI9lKwjKUpltPDvmd3KwFVekam\naO4d4+TZURo7h3i2ZQCPS1hXnseWVQVsrrLOySwr7mlCVTnWNcKjx7u5b18Hzb1jbKrM52u3X8Jr\nt1XZuKZJeC4RKguyqCzI4qq1pfgDyum+MY51DXOsa5gT3SMAPHCki+s3lnP1ujJ21BWR5U3PXr2N\nuS9DuIs/5lpur2QuVWXaF2DPt8tYAAALxUlEQVR0ysekLzjhNO0LhMY0lYCCS4IfdpdLcLsEEXAR\n/LV2xh9gxh9gfMpPRUEmp3rHONwxRNdQ8FzlHXVFvO9la3nl1ipcLlk0hzHJQlXpHpnieKjQtw9M\noIDHJVQXZ1NTlE11cTarCrMpzcs4txdBMo7PR3XMXUReDXwZcAPfVNV/nvd6JvA94FKgD3irqp5e\nbuhkoKrn1ctVVQbHZ+gemaRzcIKOwUkebuxmaGKawYkZhiZmGJ30nRtbPF8Zbhd1pTnsqCviYxsr\nuGFzORX5WVE5tjGJQkSoKghOuN6wqYLJGT+nesdo7hmlbWCC3af68Z0MfqdcAiW5GZTlZdLcM0p9\nWS5rSnOpL81hdVE2bldq/BYbtriLiBv4GnAT0A7sEZFdqnp0TrM7gAFVXS8itwGfB94ai8DRNjHt\np31gnPaBCdoHJ+ganKBraJKekSkGxqcZGJtmZMp3rvesBD8c7lCPOSM0GeR1B28et+BxCQ8dDY5f\n+wPK+LQ/ONk0OUPv6DTT/hfveOMWoTDHS2G2l4bSXPKzPORlesjN8JDldZPheWHCSQARUA1OYvkD\nwcmq2ceCkOEWvB4X2V4377t+Xcp8WI2JVJbXzZZVBWxZVQAEv4dnRybpHp6iZ2SSsyNT9I1O84Pd\nLUzOvPB9nO0MrSnLZW1o0rihNJeGshwq87Ne9Jtuoouk53450KSqzQAi8mPgFmBucb8F+Ezo/s+A\nu0RE1IExnxl/gIkZPxPTfkYmgz3hoYlgUe0ZmaJnZIozQ5N0Dk3QMTBB37wLMdyuYA+gPD+TyoIs\nNlcVhM7XdXGiewSXCP6AEggovtCsfvA83eDN51emZgL0j02HCrGQm+mmLC+HvEwvZfkZVORnUZGf\nyeqibGqKs3n4aDeuGI15W2E3Jvg9WFUYHJaZ67bLajk7MsWp3jFO941xunfs3P3/eq7nRVsPZnpc\nwSGe4hyqi7KpyM+koiCTsrxMirK9FOZ4KcjyBjtlGS4y3C5H57IiKe7VQNucx+3AFYu1UVWfiAwB\npUBvNELO9cixbj71yyP4VfEHgr3V2cI649ew59rmZQbPl60uyuaC1QXUFOcEL8AI/VmWl7loQYzV\nmHusCrsxZmkulwTPny/M4qp1pS96zR9QOgcnaOkb51TfGC29Y3QMTtAxOEFjxxD949Ms1X2dvRgr\nI/QbvdsluCT45zuuqueDN6yP6b8trmfLiMidwJ2hh6MiciKe7z+rMbJmZZzHf05vW+lfPD8vyexQ\njuU4r5+zg5Ixt2WeJ4bfjyVzPw18aOXHjmih/UiKewdQO+dxTei5hdq0i4gHKCQ4sfoiqno3cHck\nwZwmInsjmZFOJJY5fpIxt2WOn0TIHcnCYXuADSKyRkQygNuAXfPa7ALeFbp/K/CoE+PtxhhjgsL2\n3ENj6B8GfkvwVMhvq2qjiHwW2Kuqu4BvAd8XkSagn+B/AMYYYxwS0Zi7qj4APDDvuU/PuT8JvDm6\n0RyXFMNH81jm+EnG3JY5fhzP7dgVqsYYY2LHNuswxpgUZMU9RERKRORhEXk+9GfxIu3qROQhETkm\nIkdFpCG+SV+UJaLMobYFItIuInfFM+MCOcJmFpHtIvJHEWkUkUMi4sjVziLyahE5ISJNIvKJBV7P\nFJF7Q6/vdvKzMFcEuf8q9Nk9JCKPiEhEp9bFUrjMc9q9SURURBw/gyaSzCLyltDPulFE7olrQFW1\nW3Bo6gvAJ0L3PwF8fpF2jwE3he7nATmJnjn0+peBe4C7Ev3nDGwENoTurwa6gKI453QDJ4G1QAZw\nENg6r80HgW+E7t8G3Ovkz3YZuV8++7kFPuB07kgyh9rlA48TPE18Z6JnBjYA+4Hi0OOKeGa0nvsL\nbgG+G7r/XeCN8xuIyFbAo6oPA6jqqKqOxy/iS4TNDCAilwKVwENxyrWUsJlV9TlVfT50vxM4y+yu\nEPFzbtkNVZ0GZpfdmGvuv+VnwCvE+bWTw+ZW1d/P+dw+TfDaFSdF8rMG+BzBdasm4xluEZFkfi/w\nNVUdAFDVs/EMaMX9BZWq2hW6f4ZgMZxvIzAoIveJyH4R+ZfQwmpOCZtZRFzA/wE+Hs9gS4jk53yO\niFxOsGd0MtbB5llo2Y3qxdqoqg+YXXbDSZHknusO4DcxTRRe2MwicglQq6r3xzPYEiL5OW8ENorI\nkyLydGh13bhJq806ROR3QNUCL31y7gNVVRFZ6DQiD3AdsANoBe4F3k3wPP+YiELmDwIPqGp7vDqV\nUcg8e5xVwPeBd6lqYLF2ZmVE5O3ATuB6p7MsJdRB+RLB71oy8RAcmrmB4G9Hj4vINlUdjNebpw1V\n/ZPFXhORbhFZpapdoaKy0K9Q7cABfWGFzF8CVxLD4h6FzFcB14nIBwnOEWSIyKiqLjppdb6ikBkR\nKQDuBz6pqk/HKOpSorbsRpxFkhsR+ROC/9ler6pTccq2mHCZ84ELgcdCHZQqYJeI3KyqTu34E8nP\nuR3YraozwCkReY5gsd8Tj4A2LPOCuUsovAv41QJt9gBFIjI7/nsjL176ON7CZlbVt6lqnao2EBya\n+V4sC3sEwmYOLXPxC4JZfxbHbHMl67IbYXOLyA7g/wI3x3sceBFLZlbVIVUtU9WG0Of4aYLZndzK\nLZLPxy8J9toRkTKCwzTNcUvo5IxzIt0IjpU+AjwP/A4oCT2/k+DuU7PtbgIOAYeB7wAZiZ55Tvt3\n4/zZMmEzA28HZoADc27bHcj6WuA5guP9nww991mChQUgC/gp0AQ8A6x18me7jNy/A7rn/Gx3JXrm\neW0fw+GzZSL8OQvB4aSjoXpxWzzz2RWqxhiTgmxYxhhjUpAVd2OMSUFW3I0xJgVZcTfGmBRkxd0Y\nY1KQFXdjjElBVtyNI0TkMyLycRH5bOhqSUTkutDSqAdEJDu0dk+jiPxLjDI0iMiRMG1+FFoa92NR\neL93R2vJ5VD226NxLJOa0mr5AZN4dM52jcDbgP+tqj8AEJE7CV7k5A93HBHxaHDxrqgRkSrgMlVd\nH83jRkkDcDvBZZyNeQnruZu4EZFPishzIvIEsCn03HdE5FYReQ/wFuBzIvJDEdlFcC2cZxfbrCP0\nd78hIruBL4hIroh8W0SeCa3aeUuoXYOI/EFE9oVuV0cY+SGgOvSbxHUi8tjsJhEiUiYip0P33x1a\nKfRBCW5C8oU5Gf8s9G9+BrgmzM+nXER+LiJ7QrdrQs9fH8pwIPTvygf+meCaQQei8VuFST3Wczdx\nIcE15W8DthP83O0Dnp19XVW/KSLXAr/W0HoyoQXOtoc5dA1wtar6ReSfCK7v8uciUgQ8E1qh8izB\nDVYmRWQD8COCyx2Ec3Moz/ZQnqXabie4WugUcEJEvgr4gL8HLiW4HPDvCW7esJgvA/+qqk+ISB3w\nW2ALwTWBPqSqT4pIHsH1zD8BfFxVXx/Bv8OkISvuJl6uA36hoU0iQj3zaPjpnGGbVwI3i8js2vVZ\nQB3QCdwlItsBP8EFnKLtEVUdAhCRo0A9UAY8pqo9oefvDfPefwJsnfOfSEGomD8JfElEfgjcp3Fc\nvtkkLyvuJtmNzbkvwJtU9cTcBiLyGYILZV1McChypTv5+HhhKDNr3mtzl831s7Lvlgu4UlXn5/tn\nEbmf4EJVT4rIq1ZwbJNmbMzdxMvjwBtDZ8HkA2+IwXv8FvgLCXVrQ0vbQnCd9S4NbvjxDoL7X67E\naYJDLBBc4jec3cD1IlIqIl7gzWHaPwT8xeyD0G8aiMg6VT2sqp8nuNTsZmCE4DrnxizIiruJC1Xd\nR3DnqoMEt3WLxYYFnwO8wCERaQw9Bvh34F0icpBgYRxb5O+H80XgAyKyn+CQy5I0uJ3gZ4A/Ehxa\nORbmr3wE2Bk69fIo8P7Q8x8VkSMicojgUsi/IbjstF9EDtqEqlmILflrjDEpyHruxhiTgmxC1SQ8\nEfkkLx2v/qmq/mOUjv8q4PPznj6lqv8tGsdf4P1i+u8xBmxYxhhjUpINyxhjTAqy4m6MMSnIirsx\nxqQgK+7GGJOCrLgbY0wK+v8cHPdFzjgtYAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11c24efd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.distplot(gtspdf[\"diff_real_fund_est\"]) # 看误差分布展宽也明显比南方原油窄"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.05"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len([i for i in gtspdf[\"diff_real_fund_est\"][-100:] if abs(i)>0.31])/100"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "极端值情况，5%分位的误差偏移仅有0.31个百分点，也比南方原油强不少，因此国泰商品预测性更强，更适合套利，或者说套利溢价的阈值更低，因为溢价预测的误差窗口小一些。\n",
    "\n",
    "我们再看一下，预测误差的正负和实际上涨与下降之间的关系，这之前先看一下完全反号的预测"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "283 0 0.0\n"
     ]
    }
   ],
   "source": [
    "p = 0\n",
    "n = 0\n",
    "for i, row in gtspdf.iterrows():\n",
    "    r = row[\"fund_est\"] * row[\"real\"]\n",
    "    if r >= 0:\n",
    "        p += 1\n",
    "    else:\n",
    "        n += 1\n",
    "print(p, n, n / (n + p))\n",
    "# 并没有反向预测这种严重错误，这是一个好消息！"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "65 99 63 56\n"
     ]
    }
   ],
   "source": [
    "uu, ud, dd, du = 0, 0, 0, 0 \n",
    "#uu 实际上涨，real-est>0 (预测涨的少)\n",
    "# ud 预测涨的多\n",
    "# du 预测跌的多\n",
    "# dd 预测跌的少\n",
    "for i, row in gtspdf.iterrows():\n",
    "    if row[\"real\"]>=0 and row[\"diff_real_fund_est\"]>0:\n",
    "        uu += 1\n",
    "    elif row[\"real\"]>=0 and row[\"diff_real_fund_est\"]<=0:\n",
    "        ud += 1\n",
    "    elif row[\"real\"]<0 and row[\"diff_real_fund_est\"]>0:\n",
    "        du += 1\n",
    "    else:\n",
    "        dd +=1\n",
    "print(uu,ud,du,dd)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "还是比较平均的，预测涨的比实际多的情况稍多。也就是说预计涨的时候，可能预计的偏大。预计跌的时候，还是没有什么系统性误差的。\n",
    "\n",
    "说人话就是基金涨的时候跟不上持仓，跌的时候跟挺紧。。。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 华宝油气净值预估情况"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "hbyqdf = estimate_table(\n",
    "    \"20190101\",\n",
    "    \"20200305\",\n",
    "    (\"real\", \"\", {\"F162411\": 100}),\n",
    "    (\"fund_est\", \"USD/CNY\", holdings_162411_19s4),\n",
    "    (\"fund_est_2111\", \"2111\", holdings_162411_19s4),\n",
    "    _default=cache_table,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>date</th>\n",
       "      <th>real</th>\n",
       "      <th>fund_est</th>\n",
       "      <th>fund_est_2111</th>\n",
       "      <th>diff_real_fund_est</th>\n",
       "      <th>diff_real_fund_est_2111</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>263</th>\n",
       "      <td>2020-02-07</td>\n",
       "      <td>-1.873199</td>\n",
       "      <td>-1.913981</td>\n",
       "      <td>-1.174162</td>\n",
       "      <td>0.040782</td>\n",
       "      <td>-0.699037</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>264</th>\n",
       "      <td>2020-02-10</td>\n",
       "      <td>-2.055800</td>\n",
       "      <td>-2.121451</td>\n",
       "      <td>-2.493273</td>\n",
       "      <td>0.065650</td>\n",
       "      <td>0.437473</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>265</th>\n",
       "      <td>2020-02-11</td>\n",
       "      <td>0.539730</td>\n",
       "      <td>0.578453</td>\n",
       "      <td>0.264689</td>\n",
       "      <td>-0.038723</td>\n",
       "      <td>0.275041</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>266</th>\n",
       "      <td>2020-02-12</td>\n",
       "      <td>0.745601</td>\n",
       "      <td>0.792262</td>\n",
       "      <td>1.136632</td>\n",
       "      <td>-0.046661</td>\n",
       "      <td>-0.391031</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>267</th>\n",
       "      <td>2020-02-13</td>\n",
       "      <td>-0.266430</td>\n",
       "      <td>-0.265953</td>\n",
       "      <td>-0.288820</td>\n",
       "      <td>-0.000477</td>\n",
       "      <td>0.022391</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>268</th>\n",
       "      <td>2020-02-14</td>\n",
       "      <td>-0.831107</td>\n",
       "      <td>-0.931193</td>\n",
       "      <td>-0.873006</td>\n",
       "      <td>0.100086</td>\n",
       "      <td>0.041899</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>269</th>\n",
       "      <td>2020-02-17</td>\n",
       "      <td>-0.089793</td>\n",
       "      <td>-0.068726</td>\n",
       "      <td>-0.080150</td>\n",
       "      <td>-0.021068</td>\n",
       "      <td>-0.009643</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>270</th>\n",
       "      <td>2020-02-18</td>\n",
       "      <td>-0.629119</td>\n",
       "      <td>-0.674867</td>\n",
       "      <td>-0.492850</td>\n",
       "      <td>0.045748</td>\n",
       "      <td>-0.136270</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>271</th>\n",
       "      <td>2020-02-19</td>\n",
       "      <td>2.653000</td>\n",
       "      <td>2.920169</td>\n",
       "      <td>2.661412</td>\n",
       "      <td>-0.267169</td>\n",
       "      <td>-0.008412</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>272</th>\n",
       "      <td>2020-02-20</td>\n",
       "      <td>0.205580</td>\n",
       "      <td>0.234215</td>\n",
       "      <td>0.567880</td>\n",
       "      <td>-0.028635</td>\n",
       "      <td>-0.362300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>273</th>\n",
       "      <td>2020-02-21</td>\n",
       "      <td>-1.524033</td>\n",
       "      <td>-1.646148</td>\n",
       "      <td>-1.845239</td>\n",
       "      <td>0.122115</td>\n",
       "      <td>0.321206</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>274</th>\n",
       "      <td>2020-02-24</td>\n",
       "      <td>-5.089286</td>\n",
       "      <td>-5.531251</td>\n",
       "      <td>-5.548761</td>\n",
       "      <td>0.441965</td>\n",
       "      <td>0.459475</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>275</th>\n",
       "      <td>2020-02-25</td>\n",
       "      <td>-4.358733</td>\n",
       "      <td>-4.790551</td>\n",
       "      <td>-4.986971</td>\n",
       "      <td>0.431817</td>\n",
       "      <td>0.628238</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>276</th>\n",
       "      <td>2020-02-26</td>\n",
       "      <td>-4.098361</td>\n",
       "      <td>-4.569926</td>\n",
       "      <td>-4.303033</td>\n",
       "      <td>0.471565</td>\n",
       "      <td>0.204672</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>277</th>\n",
       "      <td>2020-02-27</td>\n",
       "      <td>-5.094017</td>\n",
       "      <td>-5.641765</td>\n",
       "      <td>-6.002919</td>\n",
       "      <td>0.547748</td>\n",
       "      <td>0.908902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>278</th>\n",
       "      <td>2020-02-28</td>\n",
       "      <td>2.737752</td>\n",
       "      <td>3.068003</td>\n",
       "      <td>3.102861</td>\n",
       "      <td>-0.330251</td>\n",
       "      <td>-0.365109</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>279</th>\n",
       "      <td>2020-03-02</td>\n",
       "      <td>-0.736325</td>\n",
       "      <td>-0.848291</td>\n",
       "      <td>-0.927326</td>\n",
       "      <td>0.111965</td>\n",
       "      <td>0.191001</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>280</th>\n",
       "      <td>2020-03-03</td>\n",
       "      <td>-2.825857</td>\n",
       "      <td>-3.158663</td>\n",
       "      <td>-2.733733</td>\n",
       "      <td>0.332806</td>\n",
       "      <td>-0.092123</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>281</th>\n",
       "      <td>2020-03-04</td>\n",
       "      <td>0.399855</td>\n",
       "      <td>0.450548</td>\n",
       "      <td>-0.063114</td>\n",
       "      <td>-0.050693</td>\n",
       "      <td>0.462968</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>282</th>\n",
       "      <td>2020-03-05</td>\n",
       "      <td>-3.511948</td>\n",
       "      <td>-3.924178</td>\n",
       "      <td>-3.616300</td>\n",
       "      <td>0.412230</td>\n",
       "      <td>0.104352</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          date      real  fund_est  fund_est_2111  diff_real_fund_est  \\\n",
       "263 2020-02-07 -1.873199 -1.913981      -1.174162            0.040782   \n",
       "264 2020-02-10 -2.055800 -2.121451      -2.493273            0.065650   \n",
       "265 2020-02-11  0.539730  0.578453       0.264689           -0.038723   \n",
       "266 2020-02-12  0.745601  0.792262       1.136632           -0.046661   \n",
       "267 2020-02-13 -0.266430 -0.265953      -0.288820           -0.000477   \n",
       "268 2020-02-14 -0.831107 -0.931193      -0.873006            0.100086   \n",
       "269 2020-02-17 -0.089793 -0.068726      -0.080150           -0.021068   \n",
       "270 2020-02-18 -0.629119 -0.674867      -0.492850            0.045748   \n",
       "271 2020-02-19  2.653000  2.920169       2.661412           -0.267169   \n",
       "272 2020-02-20  0.205580  0.234215       0.567880           -0.028635   \n",
       "273 2020-02-21 -1.524033 -1.646148      -1.845239            0.122115   \n",
       "274 2020-02-24 -5.089286 -5.531251      -5.548761            0.441965   \n",
       "275 2020-02-25 -4.358733 -4.790551      -4.986971            0.431817   \n",
       "276 2020-02-26 -4.098361 -4.569926      -4.303033            0.471565   \n",
       "277 2020-02-27 -5.094017 -5.641765      -6.002919            0.547748   \n",
       "278 2020-02-28  2.737752  3.068003       3.102861           -0.330251   \n",
       "279 2020-03-02 -0.736325 -0.848291      -0.927326            0.111965   \n",
       "280 2020-03-03 -2.825857 -3.158663      -2.733733            0.332806   \n",
       "281 2020-03-04  0.399855  0.450548      -0.063114           -0.050693   \n",
       "282 2020-03-05 -3.511948 -3.924178      -3.616300            0.412230   \n",
       "\n",
       "     diff_real_fund_est_2111  \n",
       "263                -0.699037  \n",
       "264                 0.437473  \n",
       "265                 0.275041  \n",
       "266                -0.391031  \n",
       "267                 0.022391  \n",
       "268                 0.041899  \n",
       "269                -0.009643  \n",
       "270                -0.136270  \n",
       "271                -0.008412  \n",
       "272                -0.362300  \n",
       "273                 0.321206  \n",
       "274                 0.459475  \n",
       "275                 0.628238  \n",
       "276                 0.204672  \n",
       "277                 0.908902  \n",
       "278                -0.365109  \n",
       "279                 0.191001  \n",
       "280                -0.092123  \n",
       "281                 0.462968  \n",
       "282                 0.104352  "
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "hbyqdf.iloc[-20:]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "这个净值预估就有点意思了，因为雪球上油气姐也有每日官方的净值预报，让我们比一下谁的更准\n",
    "要知道实在不行我们还有个持仓比例的超参可以调！\n",
    "\n",
    "下面是雪球的预估数据\n",
    "3.5 -4.06\n",
    "3.4 0.33\n",
    "3.3 -3.21\n",
    "3.2 -0.77\n",
    "2.28 3.39\n",
    "\n",
    "就拿最新五天的数据为例，四天都是我们上面的预计更准，因此可以说也是全网最准了，毕竟比官方的预估要接近真实值"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.11470246307431636, 0.25056930156434193)"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(\n",
    "    np.mean(np.abs(hbyqdf[\"diff_real_fund_est\"][-200:])),\n",
    "    np.mean(np.abs(hbyqdf[\"diff_real_fund_est_2111\"][-200:])),\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "油气的预测平均偏差是0.11个百分点强，比国泰商品的预测精度还低，同样是人民币中间价比市价准很多。\n",
    "这样看国泰商品竟然是比华宝油气这种跟踪单一指数的产品净值预测更容易，估计这颠覆了很多人的认识"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(-0.03564024234015609, -0.009798129247228005, 0.02813788155226706)"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(\n",
    "    np.mean(hbyqdf[\"diff_real_fund_est\"][-150:-100]),\n",
    "    np.mean(hbyqdf[\"diff_real_fund_est\"][-100:-50]),\n",
    "    np.mean(hbyqdf[\"diff_real_fund_est\"][-50:]),\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x11ce162e8>"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAELCAYAAAAiIMZEAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nd2OMiUFW3I0xJgb9f+erI0xemPoFAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11d1ed4e0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.distplot(hbyqdf[\"diff_real_fund_est\"], bins=40) # 误差分布涨的有点丑, 好像确实也没有国泰商品的正态"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.05"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len([i for i in hbyqdf[\"diff_real_fund_est\"][-100:] if abs(i)>0.40])/100"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "百分之五极端分位是预测偏离0.4个百分点"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "283 0 0.0\n"
     ]
    }
   ],
   "source": [
    "p = 0\n",
    "n = 0\n",
    "for i, row in hbyqdf.iterrows():\n",
    "    r = row[\"fund_est\"] * row[\"real\"]\n",
    "    if r >= 0:\n",
    "        p += 1\n",
    "    else:\n",
    "        n += 1\n",
    "print(p, n, n / (n + p))\n",
    "# 无反相预测"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "78 67 58 80\n"
     ]
    }
   ],
   "source": [
    "uu, ud, dd, du = 0, 0, 0, 0 \n",
    "#uu 实际上涨，real-est>0 (预测涨的少)\n",
    "# ud 预测涨的多\n",
    "# du 预测跌的多\n",
    "# dd 预测跌的少\n",
    "for i, row in hbyqdf.iterrows():\n",
    "    if row[\"real\"]>=0 and row[\"diff_real_fund_est\"]>0:\n",
    "        uu += 1\n",
    "    elif row[\"real\"]>=0 and row[\"diff_real_fund_est\"]<=0:\n",
    "        ud += 1\n",
    "    elif row[\"real\"]<0 and row[\"diff_real_fund_est\"]>0:\n",
    "        du += 1\n",
    "    else:\n",
    "        dd +=1\n",
    "print(uu,ud,du,dd)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "对于华宝油气，预估涨的时候往往预估得不如真实涨的多；预估跌的时候，往往预估比实际跌的少，这似乎说明华宝油气预估的仓位不足，导致预估倾向比较小，让我们调一下这个超参"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [],
   "source": [
    "hbyqdf91 = estimate_table(\n",
    "    \"20190101\",\n",
    "    \"20200305\",\n",
    "    (\"real\", \"\", {\"F162411\": 100}),\n",
    "    (\"fund_est\", \"USD/CNY\", {\"1010825\": 91}),\n",
    "    (\"fund_est_2111\", \"2111\", {\"1010825\": 91}),\n",
    "    _default=cache_table,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "72 73 62 76\n"
     ]
    }
   ],
   "source": [
    "uu, ud, dd, du = 0, 0, 0, 0 \n",
    "#uu 实际上涨，real-est>0 (预测涨的少)\n",
    "# ud 预测涨的多\n",
    "# du 预测跌的多\n",
    "# dd 预测跌的少\n",
    "for i, row in hbyqdf91.iterrows():\n",
    "    if row[\"real\"]>=0 and row[\"diff_real_fund_est\"]>0:\n",
    "        uu += 1\n",
    "    elif row[\"real\"]>=0 and row[\"diff_real_fund_est\"]<=0:\n",
    "        ud += 1\n",
    "    elif row[\"real\"]<0 and row[\"diff_real_fund_est\"]>0:\n",
    "        du += 1\n",
    "    else:\n",
    "        dd +=1\n",
    "print(uu,ud,du,dd)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.11827697989010805"
      ]
     },
     "execution_count": 90,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(np.abs(hbyqdf91[\"diff_real_fund_est\"][-200:])) # 偏差大了点，但是预测与实际涨跌偏差的正负比较对称了\n",
    "# 所以仓位90 91 自己看着办吧，差不太多"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## DAX30"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "warning: this fund has no data for rate\n"
     ]
    }
   ],
   "source": [
    "cache_data(cache_table, [\"EUR/CNY\", \"/indices/germany-30\", \"F513030\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "daxdf = estimate_table(\n",
    "    \"20190101\",\n",
    "    \"20200305\",\n",
    "    (\"real\", \"\", {\"F513030\": 100}),\n",
    "    (\"fund_est\", \"EUR/CNY\", {\"/indices/germany-30\": 97}),\n",
    "    _default=cache_table,\n",
    ")\n",
    "\n",
    "# 这里的仓位97是调超参得到的最好结果"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.03418095656058967"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(np.abs(daxdf[\"diff_real_fund_est\"][-200:]))\n",
    "# 平均 0.03 偏移，德国30跟踪还是很可以的"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>date</th>\n",
       "      <th>real</th>\n",
       "      <th>fund_est</th>\n",
       "      <th>diff_real_fund_est</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>263</th>\n",
       "      <td>2020-02-07</td>\n",
       "      <td>-0.884956</td>\n",
       "      <td>-0.878367</td>\n",
       "      <td>-0.006589</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>264</th>\n",
       "      <td>2020-02-10</td>\n",
       "      <td>-0.267857</td>\n",
       "      <td>-0.259268</td>\n",
       "      <td>-0.008590</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>265</th>\n",
       "      <td>2020-02-11</td>\n",
       "      <td>0.626679</td>\n",
       "      <td>0.580633</td>\n",
       "      <td>0.046046</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>266</th>\n",
       "      <td>2020-02-12</td>\n",
       "      <td>0.711744</td>\n",
       "      <td>0.693314</td>\n",
       "      <td>0.018430</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>267</th>\n",
       "      <td>2020-02-13</td>\n",
       "      <td>-0.353357</td>\n",
       "      <td>-0.357733</td>\n",
       "      <td>0.004376</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>268</th>\n",
       "      <td>2020-02-14</td>\n",
       "      <td>-0.265957</td>\n",
       "      <td>-0.228721</td>\n",
       "      <td>-0.037236</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>269</th>\n",
       "      <td>2020-02-17</td>\n",
       "      <td>0.177778</td>\n",
       "      <td>0.216454</td>\n",
       "      <td>-0.038676</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>270</th>\n",
       "      <td>2020-02-18</td>\n",
       "      <td>-0.709849</td>\n",
       "      <td>-0.725345</td>\n",
       "      <td>0.015496</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>271</th>\n",
       "      <td>2020-02-19</td>\n",
       "      <td>0.714924</td>\n",
       "      <td>0.663141</td>\n",
       "      <td>0.051783</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>272</th>\n",
       "      <td>2020-02-20</td>\n",
       "      <td>-0.798580</td>\n",
       "      <td>-0.712748</td>\n",
       "      <td>-0.085832</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>273</th>\n",
       "      <td>2020-02-21</td>\n",
       "      <td>-0.536673</td>\n",
       "      <td>-0.541978</td>\n",
       "      <td>0.005306</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>274</th>\n",
       "      <td>2020-02-24</td>\n",
       "      <td>-3.507194</td>\n",
       "      <td>-3.424654</td>\n",
       "      <td>-0.082540</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>275</th>\n",
       "      <td>2020-02-25</td>\n",
       "      <td>-1.770736</td>\n",
       "      <td>-1.696144</td>\n",
       "      <td>-0.074593</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>276</th>\n",
       "      <td>2020-02-26</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-0.030565</td>\n",
       "      <td>0.030565</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>277</th>\n",
       "      <td>2020-02-27</td>\n",
       "      <td>-2.941176</td>\n",
       "      <td>-2.913132</td>\n",
       "      <td>-0.028044</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>278</th>\n",
       "      <td>2020-02-28</td>\n",
       "      <td>-3.030303</td>\n",
       "      <td>-2.928281</td>\n",
       "      <td>-0.102022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>279</th>\n",
       "      <td>2020-03-02</td>\n",
       "      <td>-0.100806</td>\n",
       "      <td>-0.064356</td>\n",
       "      <td>-0.036450</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>280</th>\n",
       "      <td>2020-03-03</td>\n",
       "      <td>1.412714</td>\n",
       "      <td>1.361134</td>\n",
       "      <td>0.051580</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>281</th>\n",
       "      <td>2020-03-04</td>\n",
       "      <td>1.393035</td>\n",
       "      <td>1.425901</td>\n",
       "      <td>-0.032866</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>282</th>\n",
       "      <td>2020-03-05</td>\n",
       "      <td>-1.962709</td>\n",
       "      <td>-1.970918</td>\n",
       "      <td>0.008209</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          date      real  fund_est  diff_real_fund_est\n",
       "263 2020-02-07 -0.884956 -0.878367           -0.006589\n",
       "264 2020-02-10 -0.267857 -0.259268           -0.008590\n",
       "265 2020-02-11  0.626679  0.580633            0.046046\n",
       "266 2020-02-12  0.711744  0.693314            0.018430\n",
       "267 2020-02-13 -0.353357 -0.357733            0.004376\n",
       "268 2020-02-14 -0.265957 -0.228721           -0.037236\n",
       "269 2020-02-17  0.177778  0.216454           -0.038676\n",
       "270 2020-02-18 -0.709849 -0.725345            0.015496\n",
       "271 2020-02-19  0.714924  0.663141            0.051783\n",
       "272 2020-02-20 -0.798580 -0.712748           -0.085832\n",
       "273 2020-02-21 -0.536673 -0.541978            0.005306\n",
       "274 2020-02-24 -3.507194 -3.424654           -0.082540\n",
       "275 2020-02-25 -1.770736 -1.696144           -0.074593\n",
       "276 2020-02-26  0.000000 -0.030565            0.030565\n",
       "277 2020-02-27 -2.941176 -2.913132           -0.028044\n",
       "278 2020-02-28 -3.030303 -2.928281           -0.102022\n",
       "279 2020-03-02 -0.100806 -0.064356           -0.036450\n",
       "280 2020-03-03  1.412714  1.361134            0.051580\n",
       "281 2020-03-04  1.393035  1.425901           -0.032866\n",
       "282 2020-03-05 -1.962709 -1.970918            0.008209"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "daxdf.iloc[-20:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(-0.010051134143197249, -0.0079593039572593)"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(\n",
    "    np.mean(daxdf[\"diff_real_fund_est\"][-200:-100]),\n",
    "    np.mean(daxdf[\"diff_real_fund_est\"][-200:]),\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "76 80 48 79\n"
     ]
    }
   ],
   "source": [
    "uu, ud, dd, du = 0, 0, 0, 0 \n",
    "#uu 实际上涨，real-est>0 (预测涨的少)\n",
    "# ud 预测涨的多\n",
    "# du 预测跌的多\n",
    "# dd 预测跌的少\n",
    "for i, row in daxdf.iterrows():\n",
    "    if row[\"real\"]>=0 and row[\"diff_real_fund_est\"]>0:\n",
    "        uu += 1\n",
    "    elif row[\"real\"]>=0 and row[\"diff_real_fund_est\"]<=0:\n",
    "        ud += 1\n",
    "    elif row[\"real\"]<0 and row[\"diff_real_fund_est\"]>0:\n",
    "        du += 1\n",
    "    else:\n",
    "        dd +=1\n",
    "print(uu,ud,du,dd)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x119a12780>"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAELCAYAAADJF31HAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRHzi/wOq49UR9K6kVQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x119a1db38>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "sns.distplot(daxdf[\"diff_real_fund_est\"], bins=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}