Bankruptcy Scores (#1271)

* Refacotred and expanded analytics

* Updated tests

* Updated hugo

* Updated tests

.github/workflows/test.yml

@@ -30,7 +30,7 @@ jobs:
       - run: pip install types-pytz types-requests types-termcolor types-tabulate types-PyYAML types-python-dateutil types-setuptools
       - run: bandit -x ./tests -r . || true
       - run: black --check .
-      - run: codespell --ignore-words-list=zlot,ba,buil,coo,ether,hist,hsi,mape,navagation,operatio,pres,ser,yeld,shold --quiet-level=2 --skip=./tests,.git,*.css,*.csv,*.html,*.ini,*.ipynb,*.js,*.json,*.lock,*.scss,*.txt,*.yaml
+      - run: codespell --ignore-words-list=zlot,ba,buil,coo,ether,hist,hsi,mape,navagation,operatio,pres,ser,yeld,shold,ist --quiet-level=2 --skip=./tests,.git,*.css,*.csv,*.html,*.ini,*.ipynb,*.js,*.json,*.lock,*.scss,*.txt,*.yaml
       - run: flake8 . --count --ignore=E203,W503 --max-line-length=122 --show-source --statistics
       - run: mypy --ignore-missing-imports --exclude="/setup\.py$" .
       - run: shopt -s globstar && pyupgrade --py36-plus **/*.py

.pre-commit-config.yaml

@@ -25,7 +25,7 @@ repos:
         entry: codespell
         args:
           [
-            "--ignore-words-list=zlot,ba,buil,coo,ether,hist,hsi,mape,navagation,operatio,pres,ser,yeld,shold",
+            "--ignore-words-list=zlot,ba,buil,coo,ether,hist,hsi,mape,navagation,operatio,pres,ser,yeld,shold,ist",
             "--quiet-level=2",
             "--skip=./tests,.git,*.css,*.csv,*.html,*.ini,*.ipynb,*.js,*.json,*.lock,*.scss,*.txt,*.yaml",
           ]

gamestonk_terminal/stocks/fundamental_analysis/av_model.py

@@ -1,12 +1,13 @@
 """Alpha Vantage Model"""
 __docformat__ = "numpy"
 
-from typing import Dict, Tuple
+from typing import Dict, Tuple, List
 
 import requests
 
 from alpha_vantage.fundamentaldata import FundamentalData
 import pandas as pd
+import numpy as np
 from gamestonk_terminal.stocks.fundamental_analysis.fa_helper import clean_df_index
 from gamestonk_terminal.helper_funcs import long_number_format
 from gamestonk_terminal import config_terminal as cfg
@@ -301,7 +302,27 @@ def get_earnings(ticker: str, quarterly: bool = False) -> pd.DataFrame:
     return pd.DataFrame()
 
 
-def get_fraud_ratios(ticker: str) -> Tuple[Dict[str, float], float]:
+def df_values(df: pd.DataFrame, item: str) -> List[int]:
+    """Clean the values from the df
+
+    Parameters
+    ----------
+    df : pd.DataFrame
+        The Dataframe to use
+    item : str
+        The item to select
+
+    Returns
+    -------
+    values : List[int]
+        The values for the dataframe
+    """
+    selection = df[item]
+    values = selection.apply(lambda x: int(x) if x else 0).values
+    return values.tolist()
+
+
+def get_fraud_ratios(ticker: str) -> Tuple[Dict[str, float], float, float]:
     """Get fraud ratios based on fundamentals
 
     Parameters
@@ -326,42 +347,23 @@ def get_fraud_ratios(ticker: str) -> Tuple[Dict[str, float], float]:
     df_bs = df_bs.set_index("fiscalDateEnding").iloc[:2]
     df_is = df_is.set_index("fiscalDateEnding").iloc[:2]
 
-    ar = df_bs["currentNetReceivables"].apply(lambda x: 0 if x else int(x)).values
-    sales = df_is["totalRevenue"].apply(lambda x: 0 if x else int(x)).values
-    cogs = (
-        df_is["costofGoodsAndServicesSold"].apply(lambda x: 0 if x else int(x)).values
-    )
-    ni = df_is["netIncome"].apply(lambda x: 0 if x else int(x)).values
-    ca = df_bs["totalCurrentAssets"].apply(lambda x: 0 if x else int(x)).values
-    cl = df_bs["totalCurrentLiabilities"].apply(lambda x: 0 if x else int(x)).values
-    ppe = df_bs["propertyPlantEquipment"].apply(lambda x: 0 if x else int(x)).values
-    cash = (
-        df_bs["cashAndCashEquivalentsAtCarryingValue"]
-        .apply(lambda x: 0 if x else int(x))
-        .values
-    )
-    cash_and_sec = (
-        df_bs["cashAndShortTermInvestments"].apply(lambda x: 0 if x else int(x)).values
-    )
+    ar = df_values(df_bs, "currentNetReceivables")
+    sales = df_values(df_is, "totalRevenue")
+    cogs = df_values(df_is, "costofGoodsAndServicesSold")
+    ni = df_values(df_is, "netIncome")
+    ca = df_values(df_bs, "totalCurrentAssets")
+    cl = df_values(df_bs, "totalCurrentLiabilities")
+    ppe = df_values(df_bs, "propertyPlantEquipment")
+    cash = df_values(df_bs, "cashAndCashEquivalentsAtCarryingValue")
+    cash_and_sec = df_values(df_bs, "cashAndShortTermInvestments")
     sec = [y - x for (x, y) in zip(cash, cash_and_sec)]
-    ta = df_bs["totalAssets"].apply(lambda x: 0 if x else int(x)).values
-    dep = (
-        df_bs["accumulatedDepreciationAmortizationPPE"]
-        .apply(lambda x: 0 if x else int(x))
-        .values
-    )
-    sga = (
-        df_is["sellingGeneralAndAdministrative"]
-        .apply(lambda x: 0 if x else int(x))
-        .values
-    )
-    tl = df_bs["totalLiabilities"].apply(lambda x: 0 if x else int(x)).values
-    icfo = (
-        df_is["netIncomeFromContinuingOperations"]
-        .apply(lambda x: 0 if x else int(x))
-        .values
-    )
-    cfo = df_cf["operatingCashflow"].apply(lambda x: 0 if x else int(x)).values
+    ta = df_values(df_bs, "totalAssets")
+    dep = df_values(df_bs, "accumulatedDepreciationAmortizationPPE")
+    sga = df_values(df_is, "sellingGeneralAndAdministrative")
+    tl = df_values(df_bs, "totalLiabilities")
+    icfo = df_values(df_is, "netIncomeFromContinuingOperations")
+    cfo = df_values(df_cf, "operatingCashflow")
+
     ratios: Dict = {}
     ratios["DSRI"] = (ar[0] / sales[0]) / (ar[1] / sales[1])
     ratios["GMI"] = ((sales[1] - cogs[1]) / sales[1]) / (
@@ -392,5 +394,13 @@ def get_fraud_ratios(ticker: str) -> Tuple[Dict[str, float], float]:
         + (5.679 * (tl[0] / ta[0]))
         + (0.004 * (ca[0] / cl[0]))
     )
+    v1 = np.log(ta[0] / 1000)
+    v2 = ni[0] / ta[0]
+    v3 = cash[0] / cl[0]
 
-    return ratios, zscore
+    x = ((v1 + 0.85) * v2) - 0.85
+    y = 1 + v3
+
+    mckee = x ** 2 / (x ** 2 + y ** 2)
+
+    return ratios, zscore, mckee

gamestonk_terminal/stocks/fundamental_analysis/av_view.py

@@ -171,19 +171,18 @@ def display_fraud(ticker: str):
     ticker : str
         Fundamental analysis ticker symbol
     """
-    ratios, zscore = av_model.get_fraud_ratios(ticker)
+    ratios, zscore, mckee = av_model.get_fraud_ratios(ticker)
 
     if ratios["MSCORE"] > -1.78:
-        chanceM = "high"
+        chance_m = "high"
     elif ratios["MSCORE"] > -2.22:
-        chanceM = "moderate"
+        chance_m = "moderate"
     else:
-        chanceM = "low"
+        chance_m = "low"
 
-    if zscore < 0.5:
-        chanceZ = "high"
-    else:
-        chanceZ = "low"
+    chance_z = "high" if zscore < 0.5 else "low"
+
+    chance_mcke = "low" if mckee < 0.5 else "high"
 
     if np.isnan(ratios["MSCORE"]) or np.isnan(zscore):
         console.print("Data incomplete for this ticker. Unable to calculate risk")
@@ -196,8 +195,10 @@ def display_fraud(ticker: str):
 
     console.print(
         "\n" + "MSCORE: ",
-        f"{ratios['MSCORE']:.2f} ({chanceM} chance of fraud)",
+        f"{ratios['MSCORE']:.2f} ({chance_m} chance of fraud)",
     )
 
-    console.print("ZSCORE: ", f"{zscore:.2f} ({chanceZ} chance of bankruptcy)", "\n")
+    console.print("ZSCORE: ", f"{zscore:.2f} ({chance_z} chance of bankruptcy)", "\n")
+
+    console.print("McKee: ", f"{mckee:.2f} ({chance_mcke} chance of bankruptcy)", "\n")
     return

gamestonk_terminal/stocks/fundamental_analysis/fa_controller.py

@@ -721,6 +721,10 @@ Ticker: [/param] {self.ticker} [cmds]
                 " higher probability of default. One of the criticisms that Zmijewski made was that"
                 " other bankruptcy scoring models oversampled distressed firms and favored situations"
                 " with more complete data.[Source: YCharts]"
+                "McKee-score:\n------------------------------------------------\n"
+                "The McKee Score is a bankruptcy model used to predict a firm's bankruptcy in one year"
+                "It looks at a companie's size, profitability, and liquidity to determine the probability."
+                "This model is 80% accurate in predicting bankruptcy."
             ),
         )
         ns_parser = parse_known_args_and_warn(

tests/gamestonk_terminal/stocks/fundamental_analysis/json/test_av_model/test_get_fraud_ratios.json

@@ -1 +1 @@
-[{"DSRI": NaN, "GMI": NaN, "AQI": NaN, "SGI": NaN, "DEPI": NaN, "SGAI": NaN, "LVGI": NaN, "TATA": NaN, "MSCORE": NaN}, NaN]
+[{"DSRI": 1.1154783925070624, "GMI": 1.0013237094668648, "AQI": 0.7066050530188391, "SGI": 1.2830987061599806, "DEPI": 0.9225617507161853, "SGAI": 0.9263406938399691, "LVGI": 0.713640022276499, "TATA": -0.09743422566541382, "MSCORE": -2.545490667269407}, -1.2961310063630496, 0.05928338445454072]

tests/gamestonk_terminal/stocks/fundamental_analysis/test_av_view.py

@@ -110,7 +110,7 @@ def test_check_empty_df(func, kwargs_dict, mocked_func, mocker):
 @pytest.mark.vcr(record_mode="none")
 @pytest.mark.record_stdout
 @pytest.mark.parametrize(
-    "ratios, zscore",
+    "ratios, zscore, mckscore",
     [
         (
             {
@@ -125,15 +125,16 @@ def test_check_empty_df(func, kwargs_dict, mocked_func, mocker):
                 "MSCORE": np.nan,
             },
             np.nan,
+            np.nan,
         ),
-        ({"MSCORE": -1}, 0.1),
-        ({"AQI": 1, "MSCORE": -2}, 0.1),
-        ({"MSCORE": -3}, 1),
+        ({"MSCORE": -1}, 0.1, 0.2),
+        ({"AQI": 1, "MSCORE": -2}, 0.1, 0.2),
+        ({"MSCORE": -3}, 1, 0.2),
     ],
 )
-def test_display_fraud(mocker, ratios, zscore):
+def test_display_fraud(mocker, ratios, zscore, mckscore):
     mocker.patch(
         "gamestonk_terminal.stocks.fundamental_analysis.av_view.av_model.get_fraud_ratios",
-        return_value=(ratios, zscore),
+        return_value=(ratios, zscore, mckscore),
     )
     av_view.display_fraud(ticker="TSLA")

tests/gamestonk_terminal/stocks/fundamental_analysis/txt/test_av_view/test_display_fraud[ratios0-nan-nan].txt

@@ -0,0 +1 @@
+Data incomplete for this ticker. Unable to calculate risk

tests/gamestonk_terminal/stocks/fundamental_analysis/txt/test_av_view/test_display_fraud[ratios1-0.1-0.2].txt

@@ -0,0 +1,7 @@
+Mscore Sub Stats:
+
+MSCORE:  -1.00 (high chance of fraud)
+ZSCORE:  0.10 (high chance of bankruptcy) 
+
+McKee:  0.20 (low chance of bankruptcy) 
+

tests/gamestonk_terminal/stocks/fundamental_analysis/txt/test_av_view/test_display_fraud[ratios2-0.1-0.2].txt

@@ -0,0 +1,8 @@
+Mscore Sub Stats:
+   AQI : 1.00
+
+MSCORE:  -2.00 (moderate chance of fraud)
+ZSCORE:  0.10 (high chance of bankruptcy) 
+
+McKee:  0.20 (low chance of bankruptcy) 
+

tests/gamestonk_terminal/stocks/fundamental_analysis/txt/test_av_view/test_display_fraud[ratios3-1-0.2].txt

@@ -0,0 +1,7 @@
+Mscore Sub Stats:
+
+MSCORE:  -3.00 (low chance of fraud)
+ZSCORE:  1.00 (low chance of bankruptcy) 
+
+McKee:  0.20 (low chance of bankruptcy) 
+

website/content/stocks/fundamental_analysis/fraud/_index.md

@@ -34,6 +34,11 @@ Zscore:
 ------------------------------------------------
 The Zmijewski Score is a bankruptcy model used to predict a firm's bankruptcy in two years. The ratio uses in the Zmijewski score were determined by probit analysis (think of probit as probability unit). In this case, scores less than .5 represent a higher probability of default. One of the criticisms that Zmijewski made was that other bankruptcy scoring models oversampled distressed firms and favored situations with more complete data.[Source: YCharts]
 
+McKee Score:
+------------------------------------------------
+The McKee Score is a bankruptcy model used to predict a firm's bankruptcy in one year. It looks at a companie's size, profitability, and liquidity to determine the probability. This model is 80% accurate in predicting bankruptcy.
+[Source: McKee and Lensberg](https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.619.594&rep=rep1&type=pdf)
+
 ```
 optional arguments:
   -h, --help  show this help message