new features for models + levels

model_1h.py

@@ -67,7 +67,20 @@ model_cols = [
     'RangePct_n2',
     'OHLC4_VIX',
     'OHLC4_VIX_n1',
-    'OHLC4_VIX_n2'
+    'OHLC4_VIX_n2',
+    'OpenL1',
+    'OpenL2',
+    'OpenH1',
+    'OpenH2',
+    'L1TouchPct',
+    'L2TouchPct',
+    'H1TouchPct',
+    'H2TouchPct',
+    'L1BreakPct',
+    'L2BreakPct',
+    'H1BreakPct',
+    'H2BreakPct',
+    'GreenProbas'
 ]
 
 def walk_forward_validation_seq(df, target_column_clf, target_column_regr, num_training_rows, num_periods):
@@ -324,6 +337,52 @@ def get_data():
     data['CurrentGap'] = data['CurrentGap'].shift(-1)
     data['DayOfWeek'] = pd.to_datetime(data.index)
     data['DayOfWeek'] = data['DayOfWeek'].dt.day
+    
+    # Calculate up
+    data['up'] = 100 * (data['High'].shift(1) - data['Open'].shift(1)) / data['Close'].shift(1)
+
+    # Calculate upSD
+    data['upSD'] = data['up'].rolling(30).std(ddof=0)
+
+    # Calculate aveUp
+    data['aveUp'] = data['up'].rolling(30).mean()
+    data['H1'] = data['Open'] + (data['aveUp'] / 100) * data['Open']
+    data['H2'] = data['Open'] + ((data['aveUp'] + data['upSD']) / 100) * data['Open']
+    data['down'] = 100 * (data['Open'].shift(1) - data['Low'].shift(1)) / data['Close'].shift(1)
+    data['downSD'] = data['down'].rolling(30).std(ddof=0)
+    data['aveDown'] = data['down'].rolling(30).mean()
+    data['L1'] = data['Open'] - (data['aveDown'] / 100) * data['Open']
+    data['L2'] = data['Open'] - ((data['aveDown'] + data['upSD']) / 100) * data['Open']
+
+    data = data.assign(
+        L1Touch = lambda x: x['Low'] < x['L1'],
+        L2Touch = lambda x: x['Low'] < x['L2'],
+        H1Touch = lambda x: x['High'] > x['H1'],
+        H2Touch = lambda x: x['High'] > x['H2'],
+        L1Break = lambda x: x['Close'] < x['L1'],
+        L2Break = lambda x: x['Close'] < x['L2'],
+        H1Break = lambda x: x['Close'] > x['H1'],
+        H2Break = lambda x: x['Close'] > x['H2'],
+        OpenL1 = lambda x: x['Open'] / x['L1'],
+        OpenL2 = lambda x: x['Open'] / x['L2'],
+        OpenH1 = lambda x: x['Open'] / x['H1'],
+        OpenH2 = lambda x: x['Open'] / x['H2']
+    )
+
+    level_cols = [
+        'L1Touch',
+        'L2Touch',
+        'H1Touch',
+        'H2Touch',
+        'L1Break',
+        'L2Break',
+        'H1Break',
+        'H2Break'
+    ]
+
+    for col in level_cols:
+        data[col+'Pct'] = data[col].rolling(100).mean()
+        data[col+'Pct'] = data[col+'Pct'].shift(-1)
 
     # Intraday features
     data['CurrentHigh30'] = data['High30'].shift(-1)
@@ -349,6 +408,23 @@ def get_data():
     data['Quarter'] = data['DayOfWeek'].dt.quarter
     data['DayOfWeek'] = data['DayOfWeek'].dt.weekday
 
+    def get_quintiles(df):
+        return df.groupby(pd.qcut(df['CurrentClose30toClose'], 5))['GreenDay'].mean()
+
+    probas = []
+    for i, pct in enumerate(data['CurrentClose30toClose']):
+        try:
+            df_q = get_quintiles(data.iloc[:i])
+            for q in df_q.index:
+                if q.left <= pct <= q.right:
+                    p = df_q[q]
+        except:
+            p = None
+
+        probas.append(p)
+
+    data['GreenProbas'] = probas
+
     for rid in tqdm(release_ids, desc='Merging econ data'):
         # Get the name of the release
         n = releases[rid]['name']
@@ -396,11 +472,6 @@ def get_data():
         'CurrentClose30toClose',
         'CurrentRange30',
         'GapFill30',
-        # 'OHLC4_Trend',
-        # 'OHLC4_Trend_n1',
-        # 'OHLC4_Trend_n2',
-        # 'VIX5Day',
-        # 'VIX5Day_n1',
         'CurrentGap',
         'RangePct',
         'RangePct_n1',
@@ -408,6 +479,19 @@ def get_data():
         'OHLC4_VIX',
         'OHLC4_VIX_n1',
         'OHLC4_VIX_n2',
+        'OpenL1',
+        'OpenL2',
+        'OpenH1',
+        'OpenH2',
+        'L1TouchPct',
+        'L2TouchPct',
+        'H1TouchPct',
+        'H2TouchPct',
+        'L1BreakPct',
+        'L2BreakPct',
+        'H1BreakPct',
+        'H2BreakPct',
+        'GreenProbas',
         'Target',
         'Target_clf'
         ]]

model_30m.py

@@ -68,7 +68,20 @@ model_cols = [
     'RangePct_n2',
     'OHLC4_VIX',
     'OHLC4_VIX_n1',
-    'OHLC4_VIX_n2'
+    'OHLC4_VIX_n2',
+    'OpenL1',
+    'OpenL2',
+    'OpenH1',
+    'OpenH2',
+    'L1TouchPct',
+    'L2TouchPct',
+    'H1TouchPct',
+    'H2TouchPct',
+    'L1BreakPct',
+    'L2BreakPct',
+    'H1BreakPct',
+    'H2BreakPct',
+    'GreenProbas'
 ]
 
 def walk_forward_validation_seq(df, target_column_clf, target_column_regr, num_training_rows, num_periods):
@@ -312,17 +325,19 @@ def get_data():
     data['DayOfWeek'] = data['DayOfWeek'].dt.day
 
     # Intraday features
+    data['CurrentOpen30'] = data['Open30'].shift(-1)
     data['CurrentHigh30'] = data['High30'].shift(-1)
     data['CurrentLow30'] = data['Low30'].shift(-1)
     data['CurrentClose30'] = data['Close30'].shift(-1)
 
+
     # Open to High
     data['CurrentHigh30toClose'] = (data['CurrentHigh30'] / data['Close']) - 1
     data['CurrentLow30toClose'] = (data['CurrentLow30'] / data['Close']) - 1
     data['CurrentClose30toClose'] = (data['CurrentClose30'] / data['Close']) - 1
     data['CurrentRange30'] = (data['CurrentHigh30'] - data['CurrentLow30']) / data['Close']
     data['GapFill30'] = [low <= prev_close if gap > 0 else high >= prev_close for high, low, prev_close, gap in zip(data['CurrentHigh30'], data['CurrentLow30'], data['Close'], data['CurrentGap'])]
-
+    
     # Target -- the next day's low
     data['Target'] = (data['OHLC4'] / data['PrevClose']) - 1
     data['Target'] = data['Target'].shift(-1)
@@ -335,6 +350,83 @@ def get_data():
     data['Quarter'] = data['DayOfWeek'].dt.quarter
     data['DayOfWeek'] = data['DayOfWeek'].dt.weekday
 
+    # Calculate up
+    data['up'] = 100 * (data['High'].shift(1) - data['Open'].shift(1)) / data['Close'].shift(1)
+
+    # Calculate upSD
+    data['upSD'] = data['up'].rolling(30).std(ddof=0)
+
+    # Calculate aveUp
+    data['aveUp'] = data['up'].rolling(30).mean()
+    data['H1'] = data['Open'] + (data['aveUp'] / 100) * data['Open']
+    data['H2'] = data['Open'] + ((data['aveUp'] + data['upSD']) / 100) * data['Open']
+    data['down'] = 100 * (data['Open'].shift(1) - data['Low'].shift(1)) / data['Close'].shift(1)
+    data['downSD'] = data['down'].rolling(30).std(ddof=0)
+    data['aveDown'] = data['down'].rolling(30).mean()
+    data['L1'] = data['Open'] - (data['aveDown'] / 100) * data['Open']
+    data['L2'] = data['Open'] - ((data['aveDown'] + data['upSD']) / 100) * data['Open']
+
+    data = data.assign(
+        L1Touch = lambda x: x['Low'] < x['L1'],
+        L2Touch = lambda x: x['Low'] < x['L2'],
+        H1Touch = lambda x: x['High'] > x['H1'],
+        H2Touch = lambda x: x['High'] > x['H2'],
+        L1Break = lambda x: x['Close'] < x['L1'],
+        L2Break = lambda x: x['Close'] < x['L2'],
+        H1Break = lambda x: x['Close'] > x['H1'],
+        H2Break = lambda x: x['Close'] > x['H2'],
+        OpenL1 = lambda x: np.where(x['Open'] < x['L1'], 1, 0),
+        OpenL2 = lambda x: np.where(x['Open'] < x['L2'], 1, 0),
+        OpenH1 = lambda x: np.where(x['Open'] > x['H1'], 1, 0),
+        OpenH2 = lambda x: np.where(x['Open'] > x['H2'], 1, 0),
+        CloseL1 = lambda x: np.where(x['Close'] < x['L1'], 1, 0),
+        CloseL2 = lambda x: np.where(x['Close'] < x['L2'], 1, 0),
+        CloseH1 = lambda x: np.where(x['Close'] > x['H1'], 1, 0),
+        CloseH2 = lambda x: np.where(x['Close'] > x['H2'], 1, 0)
+    )
+
+    data['OpenL1'] = data['OpenL1'].shift(-1)
+    data['OpenL2'] = data['OpenL2'].shift(-1)
+    data['OpenH1'] = data['OpenH1'].shift(-1)
+    data['OpenH2'] = data['OpenH2'].shift(-1)
+    data['CloseL1'] = data['CloseL1'].shift(-1)
+    data['CloseL2'] = data['CloseL2'].shift(-1)
+    data['CloseH1'] = data['CloseH1'].shift(-1)
+    data['CloseH2'] = data['CloseH2'].shift(-1)
+
+    level_cols = [
+        'L1Touch',
+        'L2Touch',
+        'H1Touch',
+        'H2Touch',
+        'L1Break',
+        'L2Break',
+        'H1Break',
+        'H2Break'
+    ]
+
+    for col in level_cols:
+        data[col+'Pct'] = data[col].rolling(100).mean()
+        data[col+'Pct'] = data[col+'Pct'].shift(-1)
+
+
+    def get_quintiles(df):
+        return df.groupby(pd.qcut(df['CurrentClose30toClose'], 5))['GreenDay'].mean()
+
+    probas = []
+    for i, pct in enumerate(data['CurrentClose30toClose']):
+        try:
+            df_q = get_quintiles(data.iloc[:i])
+            for q in df_q.index:
+                if q.left <= pct <= q.right:
+                    p = df_q[q]
+        except:
+            p = None
+
+        probas.append(p)
+
+    data['GreenProbas'] = probas
+
     for rid in tqdm(release_ids, desc='Merging econ data'):
         # Get the name of the release
         n = releases[rid]['name']
@@ -389,6 +481,19 @@ def get_data():
         'OHLC4_VIX',
         'OHLC4_VIX_n1',
         'OHLC4_VIX_n2',
+        'OpenL1',
+        'OpenL2',
+        'OpenH1',
+        'OpenH2',
+        'L1TouchPct',
+        'L2TouchPct',
+        'H1TouchPct',
+        'H2TouchPct',
+        'L1BreakPct',
+        'L2BreakPct',
+        'H1BreakPct',
+        'H2BreakPct',
+        'GreenProbas',
         'Target',
         'Target_clf'
         ]]

model_90m.py

@@ -67,7 +67,20 @@ model_cols = [
     'RangePct_n2',
     'OHLC4_VIX',
     'OHLC4_VIX_n1',
-    'OHLC4_VIX_n2'
+    'OHLC4_VIX_n2',
+    'OpenL1',
+    'OpenL2',
+    'OpenH1',
+    'OpenH2',
+    'L1TouchPct',
+    'L2TouchPct',
+    'H1TouchPct',
+    'H2TouchPct',
+    'L1BreakPct',
+    'L2BreakPct',
+    'H1BreakPct',
+    'H2BreakPct',
+    'GreenProbas'
 ]
 
 def walk_forward_validation_seq(df, target_column_clf, target_column_regr, num_training_rows, num_periods):
@@ -349,6 +362,69 @@ def get_data():
     data['Quarter'] = data['DayOfWeek'].dt.quarter
     data['DayOfWeek'] = data['DayOfWeek'].dt.weekday
 
+    # Calculate up
+    data['up'] = 100 * (data['High'].shift(1) - data['Open'].shift(1)) / data['Close'].shift(1)
+
+    # Calculate upSD
+    data['upSD'] = data['up'].rolling(30).std(ddof=0)
+
+    # Calculate aveUp
+    data['aveUp'] = data['up'].rolling(30).mean()
+    data['H1'] = data['Open'] + (data['aveUp'] / 100) * data['Open']
+    data['H2'] = data['Open'] + ((data['aveUp'] + data['upSD']) / 100) * data['Open']
+    data['down'] = 100 * (data['Open'].shift(1) - data['Low'].shift(1)) / data['Close'].shift(1)
+    data['downSD'] = data['down'].rolling(30).std(ddof=0)
+    data['aveDown'] = data['down'].rolling(30).mean()
+    data['L1'] = data['Open'] - (data['aveDown'] / 100) * data['Open']
+    data['L2'] = data['Open'] - ((data['aveDown'] + data['upSD']) / 100) * data['Open']
+
+    data = data.assign(
+        L1Touch = lambda x: x['Low'] < x['L1'],
+        L2Touch = lambda x: x['Low'] < x['L2'],
+        H1Touch = lambda x: x['High'] > x['H1'],
+        H2Touch = lambda x: x['High'] > x['H2'],
+        L1Break = lambda x: x['Close'] < x['L1'],
+        L2Break = lambda x: x['Close'] < x['L2'],
+        H1Break = lambda x: x['Close'] > x['H1'],
+        H2Break = lambda x: x['Close'] > x['H2'],
+        OpenL1 = lambda x: x['Open'] / x['L1'],
+        OpenL2 = lambda x: x['Open'] / x['L2'],
+        OpenH1 = lambda x: x['Open'] / x['H1'],
+        OpenH2 = lambda x: x['Open'] / x['H2']
+    )
+
+    level_cols = [
+        'L1Touch',
+        'L2Touch',
+        'H1Touch',
+        'H2Touch',
+        'L1Break',
+        'L2Break',
+        'H1Break',
+        'H2Break'
+    ]
+
+    for col in level_cols:
+        data[col+'Pct'] = data[col].rolling(100).mean()
+        data[col+'Pct'] = data[col+'Pct'].shift(-1)
+
+    def get_quintiles(df):
+        return df.groupby(pd.qcut(df['CurrentClose30toClose'], 5))['GreenDay'].mean()
+
+    probas = []
+    for i, pct in enumerate(data['CurrentClose30toClose']):
+        try:
+            df_q = get_quintiles(data.iloc[:i])
+            for q in df_q.index:
+                if q.left <= pct <= q.right:
+                    p = df_q[q]
+        except:
+            p = None
+
+        probas.append(p)
+
+    data['GreenProbas'] = probas
+
     for rid in tqdm(release_ids, desc='Merging econ data'):
         # Get the name of the release
         n = releases[rid]['name']
@@ -396,11 +472,6 @@ def get_data():
         'CurrentClose30toClose',
         'CurrentRange30',
         'GapFill30',
-        # 'OHLC4_Trend',
-        # 'OHLC4_Trend_n1',
-        # 'OHLC4_Trend_n2',
-        # 'VIX5Day',
-        # 'VIX5Day_n1',
         'CurrentGap',
         'RangePct',
         'RangePct_n1',
@@ -408,6 +479,19 @@ def get_data():
         'OHLC4_VIX',
         'OHLC4_VIX_n1',
         'OHLC4_VIX_n2',
+        'OpenL1',
+        'OpenL2',
+        'OpenH1',
+        'OpenH2',
+        'L1TouchPct',
+        'L2TouchPct',
+        'H1TouchPct',
+        'H2TouchPct',
+        'L1BreakPct',
+        'L2BreakPct',
+        'H1BreakPct',
+        'H2BreakPct',
+        'GreenProbas',
         'Target',
         'Target_clf'
         ]]

model_day.py

@@ -61,7 +61,19 @@ model_cols = [
     'RangePct_n2',
     'OHLC4_VIX',
     'OHLC4_VIX_n1',
-    'OHLC4_VIX_n2'
+    'OHLC4_VIX_n2',
+    'OpenL1',
+    'OpenL2',
+    'OpenH1',
+    'OpenH2',
+    'L1TouchPct',
+    'L2TouchPct',
+    'H1TouchPct',
+    'H2TouchPct',
+    'L1BreakPct',
+    'L2BreakPct',
+    'H1BreakPct',
+    'H2BreakPct'
 ]
 
 def walk_forward_validation_seq(df, target_column_clf, target_column_regr, num_training_rows, num_periods):
@@ -262,6 +274,51 @@ def get_data():
     data['DayOfWeek'] = pd.to_datetime(data.index)
     data['DayOfWeek'] = data['DayOfWeek'].dt.day
 
+    # Calculate up
+    data['up'] = 100 * (data['High'].shift(1) - data['Open'].shift(1)) / data['Close'].shift(1)
+
+    # Calculate upSD
+    data['upSD'] = data['up'].rolling(30).std(ddof=0)
+
+    # Calculate aveUp
+    data['aveUp'] = data['up'].rolling(30).mean()
+    data['H1'] = data['Open'] + (data['aveUp'] / 100) * data['Open']
+    data['H2'] = data['Open'] + ((data['aveUp'] + data['upSD']) / 100) * data['Open']
+    data['down'] = 100 * (data['Open'].shift(1) - data['Low'].shift(1)) / data['Close'].shift(1)
+    data['downSD'] = data['down'].rolling(30).std(ddof=0)
+    data['aveDown'] = data['down'].rolling(30).mean()
+    data['L1'] = data['Open'] - (data['aveDown'] / 100) * data['Open']
+    data['L2'] = data['Open'] - ((data['aveDown'] + data['upSD']) / 100) * data['Open']
+
+    data = data.assign(
+        L1Touch = lambda x: x['Low'] < x['L1'],
+        L2Touch = lambda x: x['Low'] < x['L2'],
+        H1Touch = lambda x: x['High'] > x['H1'],
+        H2Touch = lambda x: x['High'] > x['H2'],
+        L1Break = lambda x: x['Close'] < x['L1'],
+        L2Break = lambda x: x['Close'] < x['L2'],
+        H1Break = lambda x: x['Close'] > x['H1'],
+        H2Break = lambda x: x['Close'] > x['H2'],
+        OpenL1 = lambda x: x['Open'] / x['L1'],
+        OpenL2 = lambda x: x['Open'] / x['L2'],
+        OpenH1 = lambda x: x['Open'] / x['H1'],
+        OpenH2 = lambda x: x['Open'] / x['H2']
+    )
+
+    level_cols = [
+        'L1Touch',
+        'L2Touch',
+        'H1Touch',
+        'H2Touch',
+        'L1Break',
+        'L2Break',
+        'H1Break',
+        'H2Break'
+    ]
+
+    for col in level_cols:
+        data[col+'Pct'] = data[col].expanding().sum() / data[col].expanding().apply(len)
+
     # Target -- the next day's low
     data['Target'] = (data['OHLC4'] / data['PrevClose']) - 1
     data['Target'] = data['Target'].shift(-1)
@@ -323,6 +380,18 @@ def get_data():
         'OHLC4_VIX',
         'OHLC4_VIX_n1',
         'OHLC4_VIX_n2',
+        'OpenL1',
+        'OpenL2',
+        'OpenH1',
+        'OpenH2',
+        'L1TouchPct',
+        'L2TouchPct',
+        'H1TouchPct',
+        'H2TouchPct',
+        'L1BreakPct',
+        'L2BreakPct',
+        'H1BreakPct',
+        'H2BreakPct',
         'Target',
         'Target_clf'
         ]]

troubleshoot_day_model.ipynb

@@ -8,7 +8,7 @@
    "source": [
     "import pandas as pd\n",
     "import numpy as np\n",
-    "from model_30m import get_data, walk_forward_validation_seq"
+    "from model_day import get_data, walk_forward_validation_seq"
    ]
   },
   {
@@ -20,11 +20,10 @@
      "name": "stderr",
      "output_type": "stream",
      "text": [
-      "getting econ tickers: 100%|██████████| 3/3 [00:00<00:00,  4.25it/s]\n",
-      "Getting release dates: 100%|██████████| 8/8 [00:01<00:00,  5.09it/s]\n",
-      "Making indicators: 100%|██████████| 8/8 [00:00<00:00, 3996.48it/s]\n",
-      "Found cached dataset text (C:/Users/WINSTON-ITX/.cache/huggingface/datasets/boomsss___text/boomsss--spx_intra-e0e5e7af8fd43022/0.0.0/cb1e9bd71a82ad27976be3b12b407850fe2837d80c22c5e03a28949843a8ace2)\n",
-      "Merging econ data: 100%|██████████| 8/8 [00:00<00:00, 999.09it/s]\n"
+      "getting econ tickers: 100%|██████████| 3/3 [00:00<00:00,  6.56it/s]\n",
+      "Getting release dates: 100%|██████████| 8/8 [00:01<00:00,  5.08it/s]\n",
+      "Making indicators: 100%|██████████| 8/8 [00:00<00:00, 2664.11it/s]\n",
+      "Merging econ data: 100%|██████████| 8/8 [00:00<00:00, 1010.95it/s]\n"
      ]
     }
    ],
@@ -34,7 +33,14 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 3,
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -46,21 +52,21 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 8,
+   "execution_count": 5,
    "metadata": {},
    "outputs": [
     {
      "name": "stderr",
      "output_type": "stream",
      "text": [
-      "LR Model: 100%|██████████| 1223/1223 [00:03<00:00, 367.50it/s]\n",
-      "d:\\Projects\\gamedayspx\\model_30m.py:83: SettingWithCopyWarning: \n",
+      "LR Model: 100%|██████████| 1195/1195 [00:03<00:00, 347.11it/s]\n",
+      "d:\\Projects\\gamedayspx\\model_day.py:88: SettingWithCopyWarning: \n",
       "A value is trying to be set on a copy of a slice from a DataFrame.\n",
       "Try using .loc[row_indexer,col_indexer] = value instead\n",
       "\n",
       "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
       "  for_merge['RegrModelOut'] = for_merge['RegrModelOut'] > 0\n",
-      "CLF Model: 100%|██████████| 1123/1123 [00:08<00:00, 135.03it/s]\n"
+      "CLF Model: 100%|██████████| 1095/1095 [00:08<00:00, 129.89it/s]\n"
      ]
     }
    ],
@@ -70,7 +76,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 9,
+   "execution_count": 6,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -161,7 +167,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 10,
+   "execution_count": 7,
    "metadata": {},
    "outputs": [
     {
@@ -185,12 +191,14 @@
        "  <thead>\n",
        "    <tr style=\"text-align: right;\">\n",
        "      <th></th>\n",
-       "      <th>PctGreen</th>\n",
-       "      <th>NumObs</th>\n",
-       "      <th>NumGreen</th>\n",
+       "      <th>True</th>\n",
+       "      <th>Predicted</th>\n",
+       "      <th>Accuracy</th>\n",
+       "      <th>HTML</th>\n",
        "    </tr>\n",
        "    <tr>\n",
-       "      <th>Predicted</th>\n",
+       "      <th>index</th>\n",
+       "      <th></th>\n",
        "      <th></th>\n",
        "      <th></th>\n",
        "      <th></th>\n",
@@ -198,75 +206,211 @@
        "  </thead>\n",
        "  <tbody>\n",
        "    <tr>\n",
-       "      <th>(0.19, 0.323]</th>\n",
-       "      <td>0.149068</td>\n",
-       "      <td>161</td>\n",
-       "      <td>24</td>\n",
+       "      <th>2023-09-08</th>\n",
+       "      <td>True</td>\n",
+       "      <td>0.757584</td>\n",
+       "      <td>✅</td>\n",
+       "      <td>&amp;#9989;</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2023-09-11</th>\n",
+       "      <td>False</td>\n",
+       "      <td>0.300308</td>\n",
+       "      <td>✅</td>\n",
+       "      <td>&amp;#9989;</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2023-09-12</th>\n",
+       "      <td>True</td>\n",
+       "      <td>0.542126</td>\n",
+       "      <td>🟨</td>\n",
+       "      <td>&amp;#11036;</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2023-09-13</th>\n",
+       "      <td>True</td>\n",
+       "      <td>0.749427</td>\n",
+       "      <td>✅</td>\n",
+       "      <td>&amp;#9989;</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2023-09-14</th>\n",
+       "      <td>False</td>\n",
+       "      <td>0.385973</td>\n",
+       "      <td>✅</td>\n",
+       "      <td>&amp;#9989;</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2023-09-15</th>\n",
+       "      <td>True</td>\n",
+       "      <td>0.326088</td>\n",
+       "      <td>❌</td>\n",
+       "      <td>&amp;#10060;</td>\n",
        "    </tr>\n",
        "    <tr>\n",
-       "      <th>(0.323, 0.423]</th>\n",
-       "      <td>0.225000</td>\n",
-       "      <td>160</td>\n",
-       "      <td>36</td>\n",
+       "      <th>2023-09-18</th>\n",
+       "      <td>False</td>\n",
+       "      <td>0.341849</td>\n",
+       "      <td>✅</td>\n",
+       "      <td>&amp;#9989;</td>\n",
        "    </tr>\n",
        "    <tr>\n",
-       "      <th>(0.423, 0.527]</th>\n",
-       "      <td>0.450000</td>\n",
-       "      <td>160</td>\n",
-       "      <td>72</td>\n",
+       "      <th>2023-09-19</th>\n",
+       "      <td>False</td>\n",
+       "      <td>0.687576</td>\n",
+       "      <td>❌</td>\n",
+       "      <td>&amp;#10060;</td>\n",
        "    </tr>\n",
        "    <tr>\n",
-       "      <th>(0.527, 0.634]</th>\n",
-       "      <td>0.602484</td>\n",
-       "      <td>161</td>\n",
-       "      <td>97</td>\n",
+       "      <th>2023-09-20</th>\n",
+       "      <td>False</td>\n",
+       "      <td>0.209737</td>\n",
+       "      <td>✅</td>\n",
+       "      <td>&amp;#9989;</td>\n",
        "    </tr>\n",
        "    <tr>\n",
-       "      <th>(0.634, 0.707]</th>\n",
-       "      <td>0.681250</td>\n",
-       "      <td>160</td>\n",
-       "      <td>109</td>\n",
+       "      <th>2023-09-21</th>\n",
+       "      <td>False</td>\n",
+       "      <td>0.477048</td>\n",
+       "      <td>🟨</td>\n",
+       "      <td>&amp;#11036;</td>\n",
        "    </tr>\n",
        "    <tr>\n",
-       "      <th>(0.707, 0.773]</th>\n",
-       "      <td>0.768750</td>\n",
-       "      <td>160</td>\n",
-       "      <td>123</td>\n",
+       "      <th>2023-09-22</th>\n",
+       "      <td>True</td>\n",
+       "      <td>0.381210</td>\n",
+       "      <td>❌</td>\n",
+       "      <td>&amp;#10060;</td>\n",
        "    </tr>\n",
        "    <tr>\n",
-       "      <th>(0.773, 0.852]</th>\n",
-       "      <td>0.838509</td>\n",
-       "      <td>161</td>\n",
-       "      <td>135</td>\n",
+       "      <th>2023-09-25</th>\n",
+       "      <td>False</td>\n",
+       "      <td>0.306969</td>\n",
+       "      <td>✅</td>\n",
+       "      <td>&amp;#9989;</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2023-09-26</th>\n",
+       "      <td>True</td>\n",
+       "      <td>0.437918</td>\n",
+       "      <td>🟨</td>\n",
+       "      <td>&amp;#11036;</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2023-09-27</th>\n",
+       "      <td>True</td>\n",
+       "      <td>0.502720</td>\n",
+       "      <td>🟨</td>\n",
+       "      <td>&amp;#11036;</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2023-09-28</th>\n",
+       "      <td>False</td>\n",
+       "      <td>0.772822</td>\n",
+       "      <td>❌</td>\n",
+       "      <td>&amp;#10060;</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2023-09-29</th>\n",
+       "      <td>True</td>\n",
+       "      <td>0.456547</td>\n",
+       "      <td>🟨</td>\n",
+       "      <td>&amp;#11036;</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2023-10-02</th>\n",
+       "      <td>False</td>\n",
+       "      <td>0.278969</td>\n",
+       "      <td>✅</td>\n",
+       "      <td>&amp;#9989;</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2023-10-03</th>\n",
+       "      <td>True</td>\n",
+       "      <td>0.597911</td>\n",
+       "      <td>🟨</td>\n",
+       "      <td>&amp;#11036;</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2023-10-04</th>\n",
+       "      <td>False</td>\n",
+       "      <td>0.504107</td>\n",
+       "      <td>🟨</td>\n",
+       "      <td>&amp;#11036;</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2023-10-05</th>\n",
+       "      <td>True</td>\n",
+       "      <td>0.399995</td>\n",
+       "      <td>❌</td>\n",
+       "      <td>&amp;#10060;</td>\n",
        "    </tr>\n",
        "  </tbody>\n",
        "</table>\n",
        "</div>"
       ],
       "text/plain": [
-       "                PctGreen  NumObs  NumGreen\n",
-       "Predicted                                 \n",
-       "(0.19, 0.323]   0.149068     161        24\n",
-       "(0.323, 0.423]  0.225000     160        36\n",
-       "(0.423, 0.527]  0.450000     160        72\n",
-       "(0.527, 0.634]  0.602484     161        97\n",
-       "(0.634, 0.707]  0.681250     160       109\n",
-       "(0.707, 0.773]  0.768750     160       123\n",
-       "(0.773, 0.852]  0.838509     161       135"
+       "             True  Predicted Accuracy      HTML\n",
+       "index                                          \n",
+       "2023-09-08   True   0.757584        ✅   &#9989;\n",
+       "2023-09-11  False   0.300308        ✅   &#9989;\n",
+       "2023-09-12   True   0.542126        🟨  &#11036;\n",
+       "2023-09-13   True   0.749427        ✅   &#9989;\n",
+       "2023-09-14  False   0.385973        ✅   &#9989;\n",
+       "2023-09-15   True   0.326088        ❌  &#10060;\n",
+       "2023-09-18  False   0.341849        ✅   &#9989;\n",
+       "2023-09-19  False   0.687576        ❌  &#10060;\n",
+       "2023-09-20  False   0.209737        ✅   &#9989;\n",
+       "2023-09-21  False   0.477048        🟨  &#11036;\n",
+       "2023-09-22   True   0.381210        ❌  &#10060;\n",
+       "2023-09-25  False   0.306969        ✅   &#9989;\n",
+       "2023-09-26   True   0.437918        🟨  &#11036;\n",
+       "2023-09-27   True   0.502720        🟨  &#11036;\n",
+       "2023-09-28  False   0.772822        ❌  &#10060;\n",
+       "2023-09-29   True   0.456547        🟨  &#11036;\n",
+       "2023-10-02  False   0.278969        ✅   &#9989;\n",
+       "2023-10-03   True   0.597911        🟨  &#11036;\n",
+       "2023-10-04  False   0.504107        🟨  &#11036;\n",
+       "2023-10-05   True   0.399995        ❌  &#10060;"
       ]
      },
-     "execution_count": 10,
+     "execution_count": 7,
      "metadata": {},
      "output_type": "execute_result"
     }
    ],
    "source": [
-    "df_probas"
+    "perf_daily.tail(20)"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 7,
+   "execution_count": 8,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "157"
+      ]
+     },
+     "execution_count": 8,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "df_probas.loc[df_probas.index[0], 'NumObs']"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
    "metadata": {},
    "outputs": [
     {
@@ -297,23 +441,23 @@
        "  <tbody>\n",
        "    <tr>\n",
        "      <th>N</th>\n",
-       "      <td>1183.00</td>\n",
-       "      <td>884.00</td>\n",
+       "      <td>1095.00</td>\n",
+       "      <td>743.00</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>ROC AUC</th>\n",
-       "      <td>0.79</td>\n",
-       "      <td>0.81</td>\n",
+       "      <td>0.74</td>\n",
+       "      <td>0.77</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>Precision</th>\n",
-       "      <td>0.72</td>\n",
-       "      <td>0.75</td>\n",
+       "      <td>0.68</td>\n",
+       "      <td>0.74</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>Recall</th>\n",
-       "      <td>0.81</td>\n",
-       "      <td>0.88</td>\n",
+       "      <td>0.75</td>\n",
+       "      <td>0.83</td>\n",
        "    </tr>\n",
        "  </tbody>\n",
        "</table>\n",
@@ -321,13 +465,13 @@
       ],
       "text/plain": [
        "               All  High Confidence\n",
-       "N          1183.00           884.00\n",
-       "ROC AUC       0.79             0.81\n",
-       "Precision     0.72             0.75\n",
-       "Recall        0.81             0.88"
+       "N          1095.00           743.00\n",
+       "ROC AUC       0.74             0.77\n",
+       "Precision     0.68             0.74\n",
+       "Recall        0.75             0.83"
       ]
      },
-     "execution_count": 7,
+     "execution_count": 9,
      "metadata": {},
      "output_type": "execute_result"
     }
@@ -336,15 +480,6 @@
     "df_performance"
    ]
   },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "perf_daily.tail(20)"
-   ]
-  },
   {
    "cell_type": "code",
    "execution_count": null,
@@ -358,7 +493,7 @@
     "\n",
     "df = plot_data.copy()\n",
     "\n",
-    "y_min = df['Low'].tail(50).min()\n",
+    "y_min = df['Low'].tail(50).min() - 50\n",
     "y_max = df['High'].tail(50).max()\n",
     "\n",
     "increasing_color = '#3399ff'  # Blue\n",
@@ -415,6 +550,22 @@
     "fig.show()\n"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "XXX"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
   {
    "cell_type": "code",
    "execution_count": null,