Commit: 0627e54111783317fc221b1997836a81e6466750 - mt5-operator (git)

Commit MetaInfo

• Revision: 0627e54111783317fc221b1997836a81e6466750 (tree)
• Time: 2023-02-11 11:48:53
• Author: kazuhiro_kondow <simauma.circus@gmai...>
• Commiter: kazuhiro_kondow

Log Message

macdについてcrossoverの角度要素を検討

Change Summary

• add: Docs/datetime_tips.py
• add: investigation_study/backtesting_fx_M1_4.py
• add: investigation_study/backtesting_fx_M1_5.py
• add: investigation_study/backtesting_fx_M1_6.py
• add: investigation_study/numpy_vector_angle_1.py
• add: investigation_study/numpy_vector_angle_2.py
• delete: mt5-operator.log.2022-12-15
• modified: src/config/logging.conf
• modified: src/config/setting.ini
• modified: src/definition.py
• modified: src/exchange_rate_info.py

Incremental Difference

--- /dev/null
+++ b/Docs/datetime_tips.py
@@ -0,0 +1,64 @@
+import MetaTrader5 as mt5
+import datetime as dt
+import pytz
+
+
+cp = 'USDJPY'
+mt5.initialize()
+mt5.copy_rates_from_pos(cp, mt5.TIMEFRAME_M1, 0, 1)
+
+# 日本時間の8:58に取得
+# array([(1675994280, 131.517, 131.518, 131.516, 131.516, 6, 3, 0)],
+#       dtype=[('time', '<i8'), ('open', '<f8'), ('high', '<f8'), ('low', '<f8'), ('close', '<f8'), ('tick_volume', '<u8'), ('spread', '<i4'), ('real_volume', '<u8')])
+
+# ローカルタイムスタンプとして扱うと２時間ずれる
+dt.datetime.fromtimestamp(1675994280)
+# datetime.datetime(2023, 2, 10, 10, 58)
+
+# UTCとして扱うとこのようになる
+dt.datetime.utcfromtimestamp(1675994280)
+# datetime.datetime(2023, 2, 10, 1, 58)
+
+# なので以下のようにして日本時間に変換する
+# 日本のtimezone
+tz_local = pytz.timezone('Etc/GMT-9')
+# OANDAサーバのtimezone
+tz_oanda = pytz.timezone('Etc/GMT-2')
+
+tz_oanda.localize(dt.datetime.utcfromtimestamp(1675994280))
+# datetime.datetime(2023, 2, 10, 1, 58, tzinfo=<StaticTzInfo 'Etc/GMT-2'>)
+
+tz_oanda.localize(dt.datetime.utcfromtimestamp(1675994280)).astimezone(tz_local)
+# datetime.datetime(2023, 2, 10, 8, 58, tzinfo=<StaticTzInfo 'Etc/GMT-9'>)
+
+# dataframeの場合は次の通り
+
+local_to = dt.datetime(
+    now.year,
+    now.month,
+    now.day,
+    0,
+    0,
+    0,
+    tzinfo=tz_local
+)
+local_from = local_to - dt.timedelta(days=30)
+
+rate_to = local_to.astimezone(tz_oanda)
+rate_from = local_from.astimezone(tz_oanda)
+
+# レートを取得
+cp = 'USDJPY'
+mt5.initialize()
+rates = mt5.copy_rates_range(cp, mt5.TIMEFRAME_M1, rate_from, rate_to)
+
+df = pd.DataFrame(rates)
+
+# timeカラムをdatetime型に変換
+df.time = pd.to_datetime(df['time'], unit='s')
+# timeカラムをindexに時系列インデックスとして割り当て
+df.index = pd.DatetimeIndex(df.time, name='Date')
+# インデックスのTimeZoneをUTCに設定
+df.index = df.index.tz_localize(dt.timezone.utc)
+# インデックスのTimeZoneをLovalに変換
+df.index = df.index.tz_convert(tz_local)

--- /dev/null
+++ b/investigation_study/backtesting_fx_M1_4.py
@@ -0,0 +1,247 @@
+# 元ネタ
+# https://qiita.com/Fujinoinvestor/items/780b138e441531860e60
+
+# crossoverにangle要素を追加してどのように結果が変わるか検証する
+# 比較対象はbacktesting_fx_M1_2.pyの結果
+# UserWarning: Searching for best of 7200 configurations.
+# Processing time: 11776.894
+# Start                     2022-10-04 03:57:00
+# End                       2022-12-30 23:54:00
+# Duration                     87 days 19:57:00
+# Exposure Time [%]                   85.284329
+# Equity Final [$]                 10644.792312
+# Equity Peak [$]                  10896.238091
+# Return [%]                           6.447923
+# Buy & Hold Return [%]               -9.405472
+# Return (Ann.) [%]                   26.162524
+# Volatility (Ann.) [%]               10.079171
+# Sharpe Ratio                         2.595702
+# Sortino Ratio                        5.504483
+# Calmar Ratio                         7.157977
+# Max. Drawdown [%]                   -3.655017
+# Avg. Drawdown [%]                   -0.110736
+# Max. Drawdown Duration       22 days 13:17:00
+# Avg. Drawdown Duration        0 days 06:13:00
+# # Trades                                  871
+# Win Rate [%]                        22.962113
+# Best Trade [%]                       0.540564
+# Worst Trade [%]                     -0.085376
+# Avg. Trade [%]                       0.010165
+# Max. Trade Duration           3 days 01:51:00
+# Avg. Trade Duration           0 days 02:23:00
+# Profit Factor                        1.167371
+# Expectancy [%]                        0.01031
+# SQN                                  1.276678
+# _strategy                 MACDCross4(n1=28...
+# _equity_curve                             ...
+# _trades                        Size  Entry...
+# MACDCross4(n1=28,n2=53,ns=15)
+
+import os
+import talib as ta
+from backtesting.lib import crossover
+from backtesting import Backtest
+from backtesting import Strategy
+import pandas as pd
+import time
+
+# 現在ディレクトリ取得
+cwd = os.getcwd()
+
+# get data
+data_path = os.path.join(
+    cwd,
+    'csv',
+    'USDJPY_M1_202210040357_202212302354.csv'
+)
+
+# csvがtab区切りなのでreat_tableを使う
+df = pd.read_table(
+    data_path
+)
+
+# DataFrameのカラム名をbacktesting向けに修正
+df = df.rename(
+    columns={
+        "<DATE>": "Date",
+        "<TIME>": "Time",
+        "<OPEN>": "Open",
+        "<HIGH>": "High",
+        "<LOW>": "Low",
+        "<CLOSE>": "Close",
+        "<TICKVOL>": "Volume",
+        "<VOL>": "not_use",
+        "<SPREAD>": "Spread",
+    }
+)
+
+# 日付に時間文字列を結合
+df.Date = df['Date'] + ' ' + df['Time']
+
+# Dateカラムをdatetime型に変換
+df.Date = pd.to_datetime(df['Date'], format='%Y.%m.%d %H:%M:%S')
+
+# indexをDateカラムに設定
+df = df.set_index('Date')
+
+# 必要なカラムのみ取得
+data = df.loc[:, ['Open', 'High', 'Low', 'Close', 'Volume']]
+
+
+def MACD(close, n1, n2, ns):
+    macd, macdsignal, macdhist = ta.MACD(
+        close, fastperiod=n1, slowperiod=n2, signalperiod=ns)
+    return macd, macdsignal
+
+
+# 下記記事を参考にStrategyを修正
+# https://mmorley.hatenablog.com/entry/fx_backtesting01
+
+class MACDCross(Strategy):
+    # 短期EMAの期間
+    n1 = 28
+    # 長期EMAの期間
+    n2 = 53
+    # シグナル（MACDのSMA）の期間
+    ns = 15
+    # angleの閾値
+    crossover_angle = 0.5
+
+    def init(self):
+        self.macd, self.macdsignal = self.I(
+            MACD, self.data.Close, self.n1, self.n2, self.ns)
+
+    # チャートデータの行ごとに呼び出される
+    # tp（take profit）利確する価格
+    # sl（stop losst）損切りする価格
+    def next(self):
+        # macdがsignalを上回った時
+        if crossover(self.macd, self.macdsignal) and self.macd[-1] < 0.0:
+            # 買いオーダー
+            self.buy(
+                tp=self.data.Close[-1] + 0.5,
+                sl=self.data.Close[-1] - 0.1
+            )
+
+        # signalがmacdを上回った時
+        elif crossover(self.macdsignal, self.macd) and self.macd[-1] > 0.0:
+            # 売りオーダー
+            self.sell(
+                tp=self.data.Close[-1] - 0.5,
+                sl=self.data.Close[-1] + 0.1
+            )
+
+# まずはStrategyが少し異なる(比較元はmacdの0以上、以下を判定してなかった)ので、その検証
+
+
+# バックテストを設定
+bt = Backtest(
+    # チャートデータ
+    data,
+    # 売買戦略
+    MACDCross,
+    # 最初の所持金
+    cash=10000,
+    # 取引手数料（為替価格に対する倍率で指定、為替価格100円でcommission=0.0005なら0.05円）
+    # OANDAの場合0.3〜0.9銭=0.003~0.009円
+    commission=0.00009,
+    # レバレッジ倍率の逆数（0.5で2倍レバレッジ）
+    margin=1.0,
+    # True：現在の終値で取引，False：次の時間の始値で取引
+    trade_on_close=True
+)
+
+s_time = time.perf_counter()
+# バックテスト実行
+output = bt.run()
+e_time = time.perf_counter()
+# 経過時間を出力(秒)
+p_time = round((e_time - s_time), 3)
+print(f'Processing time: {p_time}')
+
+# 実行結果(データ)
+print(output)
+
+# とても駄目
+# Processing time: 5.412
+# Start                     2022-10-04 03:57:00
+# End                       2022-12-30 23:54:00
+# Duration                     87 days 19:57:00
+# Exposure Time [%]                   80.680482
+# Equity Final [$]                  9033.372661
+# Equity Peak [$]                  10078.256653
+# Return [%]                          -9.666273
+# Buy & Hold Return [%]               -9.405472
+# Return (Ann.) [%]                  -33.651759
+# Volatility (Ann.) [%]                5.827159
+# Sharpe Ratio                              0.0
+# Sortino Ratio                             0.0
+# Calmar Ratio                              0.0
+# Max. Drawdown [%]                  -11.131486
+# Avg. Drawdown [%]                   -0.272737
+# Max. Drawdown Duration       85 days 20:48:00
+# Avg. Drawdown Duration        1 days 17:17:00
+# # Trades                                  822
+# Win Rate [%]                        22.262774
+# Best Trade [%]                       0.399431
+# Worst Trade [%]                     -0.085376
+# Avg. Trade [%]                      -0.002357
+# Max. Trade Duration           3 days 10:04:00
+# Avg. Trade Duration           0 days 02:35:00
+# Profit Factor                        0.964156
+# Expectancy [%]                      -0.002238
+# SQN                                 -2.511314
+# _strategy                           MACDCross
+# _equity_curve                             ...
+# _trades                        Size  Entry...
+# dtype: object
+
+# 最適化
+s_time = time.perf_counter()
+output2 = bt.optimize(
+    n1=range(20, 40, 1),
+    n2=range(40, 60, 1),
+    ns=range(2, 20, 1),
+    maximize='SQN',
+    method='grid'
+)
+e_time = time.perf_counter()
+# 経過時間を出力(秒)
+p_time = round((e_time - s_time), 3)
+print(f'Processing time: {p_time}')
+print(output2)
+print(output2._strategy)
+
+# Processing time: 36708.622
+# Start                     2022-10-04 03:57:00
+# End                       2022-12-30 23:54:00
+# Duration                     87 days 19:57:00
+# Exposure Time [%]                   78.989816
+# Equity Final [$]                  9818.920317
+# Equity Peak [$]                  10081.403521
+# Return [%]                          -1.810797
+# Buy & Hold Return [%]               -9.405472
+# Return (Ann.) [%]                   -7.448111
+# Volatility (Ann.) [%]                 7.25267
+# Sharpe Ratio                              0.0
+# Sortino Ratio                             0.0
+# Calmar Ratio                              0.0
+# Max. Drawdown [%]                   -4.752928
+# Avg. Drawdown [%]                   -0.310453
+# Max. Drawdown Duration       50 days 07:50:00
+# Avg. Drawdown Duration        2 days 10:29:00
+# # Trades                                  727
+# Win Rate [%]                        20.770289
+# Best Trade [%]                       0.399431
+# Worst Trade [%]                     -0.085338
+# Avg. Trade [%]                       0.002205
+# Max. Trade Duration           3 days 10:02:00
+# Avg. Trade Duration           0 days 02:33:00
+# Profit Factor                        1.036753
+# Expectancy [%]                       0.002341
+# SQN                                 -0.429096
+# _strategy                 MACDCross(n1=37,...
+# _equity_curve                             ...
+# _trades                        Size  Entry...
+# dtype: object
+# MACDCross(n1=37,n2=46,ns=18)

--- /dev/null
+++ b/investigation_study/backtesting_fx_M1_5.py
@@ -0,0 +1,245 @@
+# 元ネタ
+# https://qiita.com/Fujinoinvestor/items/780b138e441531860e60
+
+# 過去90日のrateを取得してbacktestingを実行する
+
+import talib as ta
+from backtesting.lib import crossover
+from backtesting import Backtest
+from backtesting import Strategy
+import pandas as pd
+import datetime as dt
+import pytz
+import MetaTrader5 as mt5
+import numpy as np
+from numpy import linalg as LA
+
+# 本日日付
+now = dt.datetime.now()
+# 日本のtimezone
+tz_local = pytz.timezone('Etc/GMT-9')
+# OANDAサーバのtimezone
+tz_oanda = pytz.timezone('Etc/GMT-2')
+
+local_to = dt.datetime(
+    now.year,
+    now.month,
+    now.day,
+    0,
+    0,
+    0,
+    tzinfo=tz_local
+)
+local_from = local_to - dt.timedelta(days=90)
+
+rate_to = local_to.astimezone(tz_oanda)
+rate_from = local_from.astimezone(tz_oanda)
+
+# レートを取得
+cp = 'USDJPY'
+mt5.initialize()
+rates = mt5.copy_rates_range(cp, mt5.TIMEFRAME_M1, rate_from, rate_to)
+
+df = pd.DataFrame(rates)
+
+# timeカラムをdatetime型に変換
+df.time = pd.to_datetime(df['time'], unit='s')
+# timeカラムをindexに時系列インデックスとして割り当て
+df.index = pd.DatetimeIndex(df.time, name='Date')
+# インデックスのTimeZoneをUTCに設定
+df.index = df.index.tz_localize(dt.timezone.utc)
+# インデックスのTimeZoneをLovalに変換
+df.index = df.index.tz_convert(tz_local)
+
+# DataFrameのカラム名をbacktesting向けに修正
+df = df.rename(
+    columns={
+        "time": "Date",
+        "open": "Open",
+        "high": "High",
+        "low": "Low",
+        "close": "Close",
+        "tick_volume": "Volume",
+        "spread": "Spread",
+        "real_volume": "not_use",
+    }
+)
+
+# 必要なカラムのみ取得
+data = df.loc[:, ['Open', 'High', 'Low', 'Close', 'Volume']]
+
+
+def MACD(close, n1, n2, ns):
+    macd, macdsignal, macdhist = ta.MACD(
+        close, fastperiod=n1, slowperiod=n2, signalperiod=ns)
+    return macd, macdsignal
+
+
+# 下記記事を参考にStrategyを修正
+# https://mmorley.hatenablog.com/entry/fx_backtesting01
+
+class MACDCross(Strategy):
+    # 短期EMAの期間
+    n1 = 28
+    # 長期EMAの期間
+    n2 = 53
+    # シグナル（MACDのSMA）の期間
+    ns = 15
+    # angleの閾値
+    crossover_angle = 0.0
+
+    def init(self):
+        self.macd, self.macdsignal = self.I(
+            MACD, self.data.Close, self.n1, self.n2, self.ns)
+
+    def vector_angle(self, u: np.ndarray, v: np.ndarray):
+        i = np.inner(u, v)
+        n = LA.norm(u) * LA.norm(v)
+        c = i / n
+        return np.rad2deg(np.arccos(np.clip(c, -1.0, 1.0)))
+
+    # チャートデータの行ごとに呼び出される
+    # tp（take profit）利確する価格
+    # sl（stop losst）損切りする価格
+    def next(self):
+        # macdがsignalを上回った時
+        if crossover(self.macd, self.macdsignal) and \
+            self.macd[-1] < 0.0 and \
+                self.vector_angle(
+                    self.macd[-2:],
+                    self.macdsignal[-2:]
+        ) > self.crossover_angle:
+            # 買いオーダー
+            self.buy(
+                tp=self.data.Close[-1] + 0.5,
+                sl=self.data.Close[-1] - 0.1
+            )
+
+        # signalがmacdを上回った時
+        elif crossover(self.macdsignal, self.macd) and \
+            self.macd[-1] > 0.0 and \
+                self.vector_angle(
+                    self.macd[-2:],
+                    self.macdsignal[-2:]
+        ) > self.crossover_angle:
+            # 売りオーダー
+            self.sell(
+                tp=self.data.Close[-1] - 0.5,
+                sl=self.data.Close[-1] + 0.1
+            )
+
+# まずはStrategyが少し異なる(比較元はmacdの0以上、以下を判定してなかった)ので、その検証
+
+
+# バックテストを設定
+bt = Backtest(
+    # チャートデータ
+    data,
+    # 売買戦略
+    MACDCross,
+    # 最初の所持金
+    cash=10000,
+    # 取引手数料（為替価格に対する倍率で指定、為替価格100円でcommission=0.0005なら0.05円）
+    # OANDAの場合0.3〜0.9銭=0.003~0.009円
+    commission=0.00009,
+    # レバレッジ倍率の逆数（0.5で2倍レバレッジ）
+    margin=1.0,
+    # True：現在の終値で取引，False：次の時間の始値で取引
+    trade_on_close=True
+)
+
+s_time = dt.datetime.timestamp(dt.datetime.utcnow())
+# バックテスト実行
+output = bt.run()
+e_time = dt.datetime.timestamp(dt.datetime.utcnow())
+# 経過時間を出力(秒)
+p_time = round((e_time - s_time), 0)
+print(f'Processing time(sec): {p_time}')
+
+# 実行結果(データ)
+print(output)
+
+# Processing time(sec): 16.0
+# Start                     2022-11-14 09:00...
+# End                       2023-02-11 00:00...
+# Duration                     88 days 15:00:00
+# Exposure Time [%]                   79.312882
+# Equity Final [$]                   9613.15298
+# Equity Peak [$]                  10130.710517
+# Return [%]                           -3.86847
+# Buy & Hold Return [%]               -5.877446
+# Return (Ann.) [%]                  -11.649176
+# Volatility (Ann.) [%]                7.002821
+# Sharpe Ratio                              0.0
+# Sortino Ratio                             0.0
+# Calmar Ratio                              0.0
+# Max. Drawdown [%]                   -8.675607
+# Avg. Drawdown [%]                   -0.269104
+# Max. Drawdown Duration       77 days 02:20:00
+# Avg. Drawdown Duration        1 days 14:35:00
+# # Trades                                  888
+# Win Rate [%]                        20.833333
+# Best Trade [%]                       0.383547
+# Worst Trade [%]                     -0.087276
+# Avg. Trade [%]                       0.002029
+# Max. Trade Duration           3 days 10:04:00
+# Avg. Trade Duration           0 days 02:09:00
+# Profit Factor                        1.032843
+# Expectancy [%]                       0.002176
+# SQN                                 -0.824697
+# _strategy                           MACDCross
+# _equity_curve                             ...
+# _trades                        Size  Entry...
+# dtype: object
+
+# 最適化して検証
+# 時間がかかりすぎるのでangleのみ最適化
+s_time = dt.datetime.timestamp(dt.datetime.utcnow())
+output2 = bt.optimize(
+    # n1=range(20, 40, 1),
+    # n2=range(40, 60, 1),
+    # ns=range(10, 30, 1),
+    crossover_angle=[0.0, 0.5, 1.0, 1.5, 2.0, 2.5, 3.0, 3.5, 4.0, 4.5, 5.0],
+    maximize='SQN',
+    method='grid'
+)
+e_time = dt.datetime.timestamp(dt.datetime.utcnow())
+# 経過時間を出力(秒)
+p_time = round((e_time - s_time), 0)
+print(f'Processing time(sec): {p_time}')
+print(output2)
+print(output2._strategy)
+
+# Processing time(sec): 132.0
+# Start                     2022-11-14 09:00...
+# End                       2023-02-11 00:00...
+# Duration                     88 days 15:00:00
+# Exposure Time [%]                   29.644473
+# Equity Final [$]                 10306.515571
+# Equity Peak [$]                  10399.114836
+# Return [%]                           3.065156
+# Buy & Hold Return [%]               -5.877446
+# Return (Ann.) [%]                   10.297861
+# Volatility (Ann.) [%]                5.746197
+# Sharpe Ratio                         1.792118
+# Sortino Ratio                        3.735854
+# Calmar Ratio                         2.387603
+# Max. Drawdown [%]                   -4.313055
+# Avg. Drawdown [%]                   -0.120288
+# Max. Drawdown Duration       65 days 17:52:00
+# Avg. Drawdown Duration        0 days 17:15:00
+# # Trades                                  312
+# Win Rate [%]                        20.512821
+# Best Trade [%]                       1.057665
+# Worst Trade [%]                      -0.08725
+# Avg. Trade [%]                       0.009701
+# Max. Trade Duration           2 days 10:13:00
+# Avg. Trade Duration           0 days 02:36:00
+# Profit Factor                        1.149254
+# Expectancy [%]                       0.009877
+# SQN                                  0.925784
+# _strategy                 MACDCross(crosso...
+# _equity_curve                             ...
+# _trades                        Size  Entry...
+# dtype: object
+# MACDCross(crossover_angle=3.5)

--- /dev/null
+++ b/investigation_study/backtesting_fx_M1_6.py
@@ -0,0 +1,246 @@
+# 元ネタ
+# https://qiita.com/Fujinoinvestor/items/780b138e441531860e60
+
+# 過去90日のrateを取得してbacktestingを実行する
+# セオリーからGoldenCross,DeadCross判定に
+# macdの0.0以上、以下を外した場合の方が
+# backtesting_fx_M1_5.pyと比較して
+# Backtestingの結果が良かった
+
+import talib as ta
+from backtesting.lib import crossover
+from backtesting import Backtest
+from backtesting import Strategy
+import pandas as pd
+import datetime as dt
+import pytz
+import MetaTrader5 as mt5
+import numpy as np
+from numpy import linalg as LA
+
+# 本日日付
+now = dt.datetime.now()
+# 日本のtimezone
+tz_local = pytz.timezone('Etc/GMT-9')
+# OANDAサーバのtimezone
+tz_oanda = pytz.timezone('Etc/GMT-2')
+
+local_to = dt.datetime(
+    now.year,
+    now.month,
+    now.day,
+    0,
+    0,
+    0,
+    tzinfo=tz_local
+)
+local_from = local_to - dt.timedelta(days=90)
+
+rate_to = local_to.astimezone(tz_oanda)
+rate_from = local_from.astimezone(tz_oanda)
+
+# レートを取得
+cp = 'USDJPY'
+mt5.initialize()
+rates = mt5.copy_rates_range(cp, mt5.TIMEFRAME_M1, rate_from, rate_to)
+
+df = pd.DataFrame(rates)
+
+# timeカラムをdatetime型に変換
+df.time = pd.to_datetime(df['time'], unit='s')
+# timeカラムをindexに時系列インデックスとして割り当て
+df.index = pd.DatetimeIndex(df.time, name='Date')
+# インデックスのTimeZoneをUTCに設定
+df.index = df.index.tz_localize(dt.timezone.utc)
+# インデックスのTimeZoneをLovalに変換
+df.index = df.index.tz_convert(tz_local)
+
+# DataFrameのカラム名をbacktesting向けに修正
+df = df.rename(
+    columns={
+        "time": "Date",
+        "open": "Open",
+        "high": "High",
+        "low": "Low",
+        "close": "Close",
+        "tick_volume": "Volume",
+        "spread": "Spread",
+        "real_volume": "not_use",
+    }
+)
+
+# 必要なカラムのみ取得
+data = df.loc[:, ['Open', 'High', 'Low', 'Close', 'Volume']]
+
+
+def MACD(close, n1, n2, ns):
+    macd, macdsignal, macdhist = ta.MACD(
+        close, fastperiod=n1, slowperiod=n2, signalperiod=ns)
+    return macd, macdsignal
+
+
+# 下記記事を参考にStrategyを修正
+# https://mmorley.hatenablog.com/entry/fx_backtesting01
+
+class MACDCross(Strategy):
+    # 短期EMAの期間
+    n1 = 28
+    # 長期EMAの期間
+    n2 = 53
+    # シグナル（MACDのSMA）の期間
+    ns = 15
+    # angleの閾値
+    crossover_angle = 0.0
+
+    def init(self):
+        self.macd, self.macdsignal = self.I(
+            MACD, self.data.Close, self.n1, self.n2, self.ns)
+
+    def vector_angle(self, u: np.ndarray, v: np.ndarray):
+        i = np.inner(u, v)
+        n = LA.norm(u) * LA.norm(v)
+        c = i / n
+        return np.rad2deg(np.arccos(np.clip(c, -1.0, 1.0)))
+
+    # チャートデータの行ごとに呼び出される
+    # tp（take profit）利確する価格
+    # sl（stop losst）損切りする価格
+    def next(self):
+        # macdがsignalを上回った時
+        if crossover(self.macd, self.macdsignal) and \
+                self.vector_angle(
+                    self.macd[-2:],
+                    self.macdsignal[-2:]
+        ) > self.crossover_angle:
+            # 買いオーダー
+            self.buy(
+                tp=self.data.Close[-1] + 0.5,
+                sl=self.data.Close[-1] - 0.1
+            )
+
+        # signalがmacdを上回った時
+        elif crossover(self.macdsignal, self.macd) and \
+                self.vector_angle(
+                    self.macd[-2:],
+                    self.macdsignal[-2:]
+        ) > self.crossover_angle:
+            # 売りオーダー
+            self.sell(
+                tp=self.data.Close[-1] - 0.5,
+                sl=self.data.Close[-1] + 0.1
+            )
+
+
+# バックテストを設定
+bt = Backtest(
+    # チャートデータ
+    data,
+    # 売買戦略
+    MACDCross,
+    # 最初の所持金
+    cash=10000,
+    # 取引手数料（為替価格に対する倍率で指定、為替価格100円でcommission=0.0005なら0.05円）
+    # OANDAの場合0.3〜0.9銭=0.003~0.009円
+    commission=0.00009,
+    # レバレッジ倍率の逆数（0.5で2倍レバレッジ）
+    margin=1.0,
+    # True：現在の終値で取引，False：次の時間の始値で取引
+    trade_on_close=True
+)
+
+s_time = dt.datetime.timestamp(dt.datetime.utcnow())
+# バックテスト実行
+output = bt.run()
+e_time = dt.datetime.timestamp(dt.datetime.utcnow())
+# 経過時間を出力(秒)
+p_time = round((e_time - s_time), 0)
+print(f'Processing time(sec): {p_time}')
+
+# 実行結果(データ)
+print(output)
+
+# Processing time(sec): 20.0
+# Start                     2022-11-14 09:00...
+# End                       2023-02-11 00:00...
+# Duration                     88 days 15:00:00
+# Exposure Time [%]                   84.418508
+# Equity Final [$]                 10681.733506
+# Equity Peak [$]                  10725.062947
+# Return [%]                           6.817335
+# Buy & Hold Return [%]               -5.877446
+# Return (Ann.) [%]                    23.66191
+# Volatility (Ann.) [%]                9.973056
+# Sharpe Ratio                         2.372584
+# Sortino Ratio                        4.343607
+# Calmar Ratio                         6.480834
+# Max. Drawdown [%]                    -3.65106
+# Avg. Drawdown [%]                   -0.124067
+# Max. Drawdown Duration       42 days 14:15:00
+# Avg. Drawdown Duration        0 days 07:42:00
+# # Trades                                 1041
+# Win Rate [%]                        24.207493
+# Best Trade [%]                       1.057665
+# Worst Trade [%]                     -0.087407
+# Avg. Trade [%]                       0.012589
+# Max. Trade Duration           2 days 08:09:00
+# Avg. Trade Duration           0 days 02:07:00
+# Profit Factor                        1.202623
+# Expectancy [%]                       0.012749
+# SQN                                  1.205884
+# _strategy                           MACDCross
+# _equity_curve                             ...
+# _trades                         Size  Entr...
+# dtype: object
+
+
+# 最適化して検証
+# 時間がかかりすぎるのでangleのみ最適化
+s_time = dt.datetime.timestamp(dt.datetime.utcnow())
+output2 = bt.optimize(
+    # n1=range(20, 40, 1),
+    # n2=range(40, 60, 1),
+    # ns=range(10, 30, 1),
+    crossover_angle=[0.0, 0.5, 1.0, 1.5, 2.0, 2.5, 3.0, 3.5, 4.0, 4.5, 5.0],
+    maximize='SQN',
+    method='grid'
+)
+e_time = dt.datetime.timestamp(dt.datetime.utcnow())
+# 経過時間を出力(秒)
+p_time = round((e_time - s_time), 0)
+print(f'Processing time(sec): {p_time}')
+print(output2)
+print(output2._strategy)
+
+# Processing time(sec): 163.0
+# Start                     2022-11-14 09:00...
+# End                       2023-02-11 00:00...
+# Duration                     88 days 15:00:00
+# Exposure Time [%]                   48.076332
+# Equity Final [$]                 10675.172967
+# Equity Peak [$]                  10707.548153
+# Return [%]                            6.75173
+# Buy & Hold Return [%]               -5.877446
+# Return (Ann.) [%]                   22.986772
+# Volatility (Ann.) [%]                7.923884
+# Sharpe Ratio                         2.900947
+# Sortino Ratio                        7.140729
+# Calmar Ratio                         6.643383
+# Max. Drawdown [%]                     -3.4601
+# Avg. Drawdown [%]                   -0.112675
+# Max. Drawdown Duration       44 days 09:11:00
+# Avg. Drawdown Duration        0 days 09:08:00
+# # Trades                                  495
+# Win Rate [%]                        21.212121
+# Best Trade [%]                       1.057665
+# Worst Trade [%]                     -0.087566
+# Avg. Trade [%]                       0.012992
+# Max. Trade Duration           2 days 06:53:00
+# Avg. Trade Duration           0 days 02:11:00
+# Profit Factor                        1.200226
+# Expectancy [%]                        0.01317
+# SQN                                  1.600122
+# _strategy                 MACDCross(crosso...
+# _equity_curve                             ...
+# _trades                        Size  Entry...
+# dtype: object
+# MACDCross(crossover_angle=3.5)

--- /dev/null
+++ b/investigation_study/numpy_vector_angle_1.py
@@ -0,0 +1,47 @@
+# vector angle
+# ベクトルのなす角を用いてMACDのCrossOver判定について改良を試みる
+# 参考文献
+# https://hiraocafe.com/note/2lineanglevec.html
+# https://www.mathpython.com/numpy-vector-angle
+import numpy as np
+from numpy import linalg as LA
+
+
+def vector_angle(u: np.array, v: np.array):
+    i = np.inner(u, v)
+    n = LA.norm(u) * LA.norm(v)
+    c = i / n
+    return np.rad2deg(np.arccos(np.clip(c, -1.0, 1.0)))
+
+
+u = np.array([3, 4])
+v = np.array([-4, 3])
+print(vector_angle(u, v))  # 90.0
+
+# テストケース
+# 2023-02-03 13:24:02,027
+# num[-2]: MACD: 0.002432, MACDSignal: 0.003394
+# num[-1]: MACD: 0.00243, MACDSignal: 0.003274
+# Technical Indicator: Consider it a DeadCross.
+u1 = np.array([0.00243, 0.002432])
+v1 = np.array([0.003274, 0.003394])
+print(vector_angle(u1, v1))
+# 1.0074377034490547
+
+# 2023-02-03 13:25:02,727
+# num[-2]: MACD: 0.002782, MACDSignal: 0.003317
+# num[-1]: MACD: 0.003056, MACDSignal: 0.003285
+# Technical Indicator: Consider it a DeadCross.
+u2 = np.array([0.003056, 0.002782])
+v2 = np.array([0.003285, 0.003317])
+print(vector_angle(u2, v2))
+# 2.964852801558037
+
+# 2023-02-03 13:42:02,638
+# num[-2]: MACD: 0.007748, MACDSignal: 0.007773
+# num[-1]: MACD: 0.006698, MACDSignal: 0.007639
+# Technical Indicator: Consider it a DeadCross.
+u3 = np.array([0.007748, 0.007773])
+v3 = np.array([0.006698, 0.007639])
+print(vector_angle(u3, v3))
+# 3.6628990024599957

--- /dev/null
+++ b/investigation_study/numpy_vector_angle_2.py
@@ -0,0 +1,120 @@
+# vector angle
+# ベクトルのなす角を用いてMACDのCrossOver判定について改良を試みる
+# 参考文献
+# https://hiraocafe.com/note/2lineanglevec.html
+# https://www.mathpython.com/numpy-vector-angle
+# 過去データ検証
+import MetaTrader5 as mt5
+import talib as ta
+import numpy as np
+from datetime import datetime
+from datetime import timedelta
+import pytz
+from numpy import linalg as LA
+
+
+def vector_angle(u: np.ndarray, v: np.ndarray):
+    i = np.inner(u, v)
+    n = LA.norm(u) * LA.norm(v)
+    c = i / n
+    return np.rad2deg(np.arccos(np.clip(c, -1.0, 1.0)))
+
+
+def crossover(series1: np.ndarray, series2: np.ndarray) -> bool:
+    """指標値列の交差判定
+    """
+    return series1[-2] < series2[-2] and series1[-1] > series2[-1]
+
+
+mt5.initialize()
+cp = 'USDJPY'
+# 短期EMAの期間
+n1 = 28
+# 長期EMAの期間
+n2 = 53
+# シグナル（MACDのSMA）の期間
+ns = 15
+
+
+# OANDAサーバのtimezone
+tz_oanda = pytz.timezone('Etc/GMT-2')
+oanda_from = datetime(2023, 2, 8, 0, 0, 0, tzinfo=tz_oanda)
+oanda_to = oanda_from + timedelta(hours=24)
+
+# レートを取得
+rates = mt5.copy_rates_range(cp, mt5.TIMEFRAME_M1, oanda_from, oanda_to)
+
+closes = rates['close']
+
+macd, macdsignal, _ = ta.MACD(
+    closes,
+    fastperiod=n1,
+    slowperiod=n2,
+    signalperiod=ns
+)
+
+macd = np.around(macd, 6)
+macdsignal = np.around(macdsignal, 6)
+
+for i in range(0, (len(rates) - 1)):
+    if np.isnan(macd[i]) or np.isnan(macd[i + 1]):
+        # データがnanは何もしない
+        continue
+    if np.isnan(macdsignal[i]) or np.isnan(macdsignal[i + 1]):
+        # データがnanは何もしない
+        continue
+    macd_s = macd[i: i + 2]
+    macdsignal_s = macdsignal[i: i + 2]
+    if crossover(macd_s, macdsignal_s) and macd_s[-1] < 0.0:
+        # GoldenCross 上昇転換サイン
+        msg = 'Technical Indicator: GoldenCross.'
+        rate_time = tz_oanda.localize(
+            datetime.utcfromtimestamp(rates[i]['time'])
+        )
+        angle = round(vector_angle(macd_s, macdsignal_s), 6)
+        print(f'{rate_time}: {angle}: {msg}')
+    elif crossover(macdsignal_s, macd_s) and macd_s[-1] > 0.0:
+        # DeadCross 降下転換サイン
+        msg = 'Technical Indicator: DeadCross.'
+        rate_time = tz_oanda.localize(
+            datetime.utcfromtimestamp(rates[i]['time'])
+        )
+        angle = round(vector_angle(macdsignal_s, macd_s), 6)
+        print(f'{rate_time}: {angle}: {msg}')
+
+# angleが0.5以上を採用すれば誤判定を少しへらせるかも
+# backtestで検証する
+
+# 2023-02-07 23:46:00+02:00: 0.48448: Technical Indicator: GoldenCross.
+# 2023-02-07 23:50:00+02:00: 1.119538: Technical Indicator: GoldenCross.
+# 2023-02-08 00:18:00+02:00: 3.135568: Technical Indicator: GoldenCross.
+# 2023-02-08 01:00:00+02:00: 22.266663: Technical Indicator: GoldenCross.
+# 2023-02-08 01:28:00+02:00: 0.558075: Technical Indicator: DeadCross.
+# 2023-02-08 02:50:00+02:00: 1.175004: Technical Indicator: GoldenCross.
+# 2023-02-08 04:15:00+02:00: 0.211562: Technical Indicator: DeadCross.
+# 2023-02-08 04:28:00+02:00: 0.247605: Technical Indicator: DeadCross.
+# 2023-02-08 05:36:00+02:00: 1.632956: Technical Indicator: GoldenCross.
+# 2023-02-08 05:59:00+02:00: 3.524362: Technical Indicator: GoldenCross.
+# 2023-02-08 06:15:00+02:00: 0.855166: Technical Indicator: GoldenCross.
+# 2023-02-08 06:59:00+02:00: 0.940879: Technical Indicator: DeadCross.
+# 2023-02-08 07:16:00+02:00: 2.399271: Technical Indicator: DeadCross.
+# 2023-02-08 08:11:00+02:00: 1.337743: Technical Indicator: GoldenCross.
+# 2023-02-08 08:43:00+02:00: 3.563947: Technical Indicator: DeadCross.
+# 2023-02-08 09:22:00+02:00: 2.031104: Technical Indicator: DeadCross.
+# 2023-02-08 10:08:00+02:00: 0.784882: Technical Indicator: GoldenCross.
+# 2023-02-08 10:28:00+02:00: 0.961372: Technical Indicator: GoldenCross.
+# 2023-02-08 11:25:00+02:00: 0.760229: Technical Indicator: GoldenCross.
+# 2023-02-08 12:00:00+02:00: 1.235069: Technical Indicator: GoldenCross.
+# 2023-02-08 12:41:00+02:00: 1.106239: Technical Indicator: DeadCross.
+# 2023-02-08 13:49:00+02:00: 2.006404: Technical Indicator: GoldenCross.
+# 2023-02-08 14:25:00+02:00: 10.711557: Technical Indicator: DeadCross.
+# 2023-02-08 15:34:00+02:00: 0.568802: Technical Indicator: DeadCross.
+# 2023-02-08 17:03:00+02:00: 1.336369: Technical Indicator: DeadCross.
+# 2023-02-08 17:15:00+02:00: 0.909583: Technical Indicator: DeadCross.
+# 2023-02-08 17:59:00+02:00: 0.832841: Technical Indicator: DeadCross.
+# 2023-02-08 18:27:00+02:00: 4.131851: Technical Indicator: GoldenCross.
+# 2023-02-08 19:06:00+02:00: 1.267079: Technical Indicator: GoldenCross.
+# 2023-02-08 19:32:00+02:00: 1.267604: Technical Indicator: GoldenCross.
+# 2023-02-08 20:17:00+02:00: 2.140756: Technical Indicator: GoldenCross.
+# 2023-02-08 20:55:00+02:00: 2.290568: Technical Indicator: DeadCross.
+# 2023-02-08 21:59:00+02:00: 0.812178: Technical Indicator: DeadCross.

--- a/mt5-operator.log.2022-12-15
+++ /dev/null
@@ -1,71 +0,0 @@
-2022-12-15 04:59:45,523 8028 MainThread __main__:75 main [DEBUG]: set schedule.
-2022-12-15 04:59:45,528 8028 MainThread __main__:81 main [DEBUG]: start loop.
-2022-12-15 05:00:00,636 8028 MainThread __main__:33 job [DEBUG]: rate_M1 time: 2022-12-14 11:00:00 Open: 135.303 High: 135.385 Low: 135.303 Close: 135.358 Volume: 313 Spread: 3
-2022-12-15 05:00:16,666 8028 MainThread __main__:59 check_ticks [DEBUG]:                    time      bid      ask  last  volume       time_msc  flags  volume_real
-0   2022-12-14 11:00:15  135.355  135.363   0.0       0  1671015615341      4          0.0
-1   2022-12-14 11:00:15  135.361  135.368   0.0       0  1671015615492    134          0.0
-2   2022-12-14 11:00:15  135.360  135.368   0.0       0  1671015615559    130          0.0
-3   2022-12-14 11:00:15  135.358  135.366   0.0       0  1671015615618    134          0.0
-4   2022-12-14 11:00:16  135.361  135.367   0.0       0  1671015616267    134          0.0
-..                  ...      ...      ...   ...     ...            ...    ...          ...
-995 2022-12-14 11:03:50  135.327  135.332   0.0       0  1671015830632    134          0.0
-996 2022-12-14 11:03:50  135.326  135.332   0.0       0  1671015830731    130          0.0
-997 2022-12-14 11:03:52  135.326  135.331   0.0       0  1671015832092      4          0.0
-998 2022-12-14 11:03:52  135.325  135.331   0.0       0  1671015832200    130          0.0
-999 2022-12-14 11:03:52  135.324  135.331   0.0       0  1671015832526    130          0.0
-
-[1000 rows x 8 columns]
-2022-12-15 05:00:30,883 8028 MainThread __main__:33 job [DEBUG]: rate_M1 time: 2022-12-14 11:00:00 Open: 135.303 High: 135.385 Low: 135.303 Close: 135.358 Volume: 313 Spread: 3
-2022-12-15 05:00:45,974 8028 MainThread __main__:33 job [DEBUG]: rate_M1 time: 2022-12-14 11:00:00 Open: 135.303 High: 135.385 Low: 135.303 Close: 135.358 Volume: 313 Spread: 3
-2022-12-15 05:01:00,046 8028 MainThread __main__:33 job [DEBUG]: rate_M1 time: 2022-12-14 11:01:00 Open: 135.357 High: 135.358 Low: 135.326 Close: 135.347 Volume: 211 Spread: 3
-2022-12-15 05:01:15,161 8028 MainThread __main__:59 check_ticks [DEBUG]:                    time      bid      ask  last  volume       time_msc  flags  volume_real
-0   2022-12-14 11:01:15  135.332  135.341   0.0       0  1671015675064      4          0.0
-1   2022-12-14 11:01:15  135.333  135.341   0.0       0  1671015675132    130          0.0
-2   2022-12-14 11:01:15  135.332  135.340   0.0       0  1671015675191    134          0.0
-3   2022-12-14 11:01:15  135.333  135.340   0.0       0  1671015675353    130          0.0
-4   2022-12-14 11:01:15  135.335  135.341   0.0       0  1671015675838    134          0.0
-..                  ...      ...      ...   ...     ...            ...    ...          ...
-995 2022-12-14 11:05:57  135.318  135.325   0.0       0  1671015957797    134          0.0
-996 2022-12-14 11:05:57  135.318  135.324   0.0       0  1671015957870      4          0.0
-997 2022-12-14 11:05:58  135.319  135.326   0.0       0  1671015958179    134          0.0
-998 2022-12-14 11:05:58  135.319  135.325   0.0       0  1671015958290      4          0.0
-999 2022-12-14 11:05:58  135.322  135.329   0.0       0  1671015958480    134          0.0
-
-[1000 rows x 8 columns]
-2022-12-15 05:05:49,913 7540 MainThread __main__:75 main [DEBUG]: set schedule.
-2022-12-15 05:05:49,915 7540 MainThread __main__:81 main [DEBUG]: start loop.
-2022-12-15 05:06:00,946 7540 MainThread __main__:33 job [DEBUG]: rate_M1 time: 2022-12-14 11:06:00 Open: 135.317 High: 135.322 Low: 135.277 Close: 135.277 Volume: 124 Spread: 3
-2022-12-15 05:06:15,978 7540 MainThread __main__:59 check_ticks [DEBUG]:                    time      bid      ask  last  volume       time_msc  flags  volume_real
-0   2022-12-14 11:06:15  135.302  135.309   0.0       0  1671015975199    134          0.0
-1   2022-12-14 11:06:15  135.302  135.306   0.0       0  1671015975250      4          0.0
-2   2022-12-14 11:06:15  135.302  135.309   0.0       0  1671015975309      4          0.0
-3   2022-12-14 11:06:15  135.301  135.308   0.0       0  1671015975364    134          0.0
-4   2022-12-14 11:06:15  135.300  135.306   0.0       0  1671015975459    134          0.0
-..                  ...      ...      ...   ...     ...            ...    ...          ...
-995 2022-12-14 11:12:31  135.226  135.230   0.0       0  1671016351385    134          0.0
-996 2022-12-14 11:12:31  135.223  135.230   0.0       0  1671016351436    130          0.0
-997 2022-12-14 11:12:31  135.224  135.229   0.0       0  1671016351490    134          0.0
-998 2022-12-14 11:12:31  135.220  135.227   0.0       0  1671016351541    134          0.0
-999 2022-12-14 11:12:31  135.220  135.225   0.0       0  1671016351759      4          0.0
-
-[1000 rows x 8 columns]
-2022-12-15 05:06:30,201 7540 MainThread __main__:33 job [DEBUG]: rate_M1 time: 2022-12-14 11:06:00 Open: 135.317 High: 135.322 Low: 135.277 Close: 135.277 Volume: 124 Spread: 3
-2022-12-15 05:06:45,296 7540 MainThread __main__:33 job [DEBUG]: rate_M1 time: 2022-12-14 11:06:00 Open: 135.317 High: 135.322 Low: 135.277 Close: 135.277 Volume: 124 Spread: 3
-2022-12-15 05:08:40,838 972 MainThread __main__:76 main [DEBUG]: set schedule.
-2022-12-15 05:08:40,843 972 MainThread __main__:82 main [DEBUG]: start loop.
-2022-12-15 05:08:45,867 972 MainThread __main__:33 job [DEBUG]: rate_M1 time: 2022-12-14 11:08:00 Open: 135.275 High: 135.296 Low: 135.259 Close: 135.284 Volume: 157 Spread: 3
-2022-12-15 05:09:00,923 972 MainThread __main__:33 job [DEBUG]: rate_M1 time: 2022-12-14 11:09:00 Open: 135.283 High: 135.285 Low: 135.256 Close: 135.271 Volume: 140 Spread: 3
-2022-12-15 05:09:15,972 972 MainThread __main__:60 check_ticks [DEBUG]: time
-2022-12-15 05:09:15,972 972 MainThread __main__:60 check_ticks [DEBUG]: bid
-2022-12-15 05:09:15,972 972 MainThread __main__:60 check_ticks [DEBUG]: ask
-2022-12-15 05:09:15,981 972 MainThread __main__:60 check_ticks [DEBUG]: last
-2022-12-15 05:09:15,981 972 MainThread __main__:60 check_ticks [DEBUG]: volume
-2022-12-15 05:09:15,982 972 MainThread __main__:60 check_ticks [DEBUG]: time_msc
-2022-12-15 05:09:15,982 972 MainThread __main__:60 check_ticks [DEBUG]: flags
-2022-12-15 05:09:15,983 972 MainThread __main__:60 check_ticks [DEBUG]: volume_real
-2022-12-15 05:09:48,869 2512 MainThread __main__:75 main [DEBUG]: set schedule.
-2022-12-15 05:09:48,875 2512 MainThread __main__:81 main [DEBUG]: start loop.
-2022-12-15 05:10:00,907 2512 MainThread __main__:33 job [DEBUG]: rate_M1 time: 2022-12-14 11:10:00 Open: 135.271 High: 135.275 Low: 135.225 Close: 135.231 Volume: 131 Spread: 3
-2022-12-15 05:10:15,947 2512 MainThread __main__:59 check_ticks [DEBUG]: Empty DataFrame
-Columns: [time, bid, ask, last, volume, time_msc, flags, volume_real]
-Index: []

--- a/src/config/logging.conf
+++ b/src/config/logging.conf
@@ -18,7 +18,7 @@ args=(sys.stdout,)
 class=logging.handlers.TimedRotatingFileHandler
 level=DEBUG
 formatter=simple
-args=('./mt5-operator.log', 'D', 1, 10, 'utf-8')
+args=('./mt5-operator.log', 'D', 10, 10, 'utf-8')
 
 [formatters]
 keys=simple

--- a/src/config/setting.ini
+++ b/src/config/setting.ini
@@ -36,9 +36,9 @@ Allowablespread = 0.01
 # 1銭は0.01円
 OpenMergin = 0.007
 # 利益確定(円)
-ProfitTaking = 1.0
+ProfitTaking = 1.5
 # 損切(円)
-LossCut = 0.5
+LossCut = 0.2
 
 # テクニカル分析
 [Technical]

--- a/src/definition.py
+++ b/src/definition.py
@@ -2,6 +2,7 @@
 # テクニカル分析
 TECHNICAL_UNCHANGED = 0
 TECHNICAL_GOLDEN_CROSS = 1
-TECHNICAL_UPTREND = 2
-TECHNICAL_DEAD_CROSS = 3
+TECHNICAL_DEAD_CROSS = 2
+TECHNICAL_UPTREND = 3
 TECHNICAL_DOWNTREND = 4
+

--- a/src/exchange_rate_info.py
+++ b/src/exchange_rate_info.py
@@ -163,7 +163,6 @@ class ExchangeRateInfo():
                 count
             )
 
-            # for debug
             # MACDを算出
             macd, macdsignal, _ = ta.MACD(
                 closes,
@@ -172,9 +171,9 @@ class ExchangeRateInfo():
                 signalperiod=self.__ns
             )
 
-            # MACDの結果を端数処理(少数点以下5)
-            macd = np.around(macd, 5)
-            macdsignal = np.around(macdsignal, 5)
+            # MACDの結果を端数処理(少数点以下6)
+            macd = np.around(macd, 6)
+            macdsignal = np.around(macdsignal, 6)
 
             # for debug
             for i in range(-4, 0, 1):
@@ -183,28 +182,6 @@ class ExchangeRateInfo():
                 )
                 logger.debug(msg)
 
-            # 最新bidを暫定終わり値としてclosesに追加
-            # closes = np.append(closes, self.__get_latest_bid())
-
-            # MACDを算出
-            # macd, macdsignal, _ = ta.MACD(
-            #     closes,
-            #     fastperiod=self.__n1,
-            #     slowperiod=self.__n2,
-            #     signalperiod=self.__ns
-            # )
-
-            # MACDの結果を端数処理(少数点以下4)
-            # macd = np.around(macd, 4)
-            # macdsignal = np.around(macdsignal, 4)
-
-            # asum_macd = macd[-1]
-            # asum_macdsig = macdsignal[-1]
-            # msg = (
-            #     f'Assumption MACD: {asum_macd}, Signal: {asum_macdsig} .'
-            # )
-            # logger.debug(msg)
-
             # 指標判定
             if self.__crossover(macd, macdsignal) and macd[-1] < 0.0:
                 # GoldenCross 上昇転換サイン
@@ -214,22 +191,24 @@ class ExchangeRateInfo():
                 # DeadCross 降下転換サイン
                 s.technical_indicator = DEF.TECHNICAL_DEAD_CROSS
                 msg = 'Technical Indicator: DeadCross.'
-            elif macd[-1] > macdsignal[-1] and macd[-1] < 0.0:
-                # 弱い上昇トレンド
-                s.technical_indicator = DEF.TECHNICAL_UPTREND
-                msg = 'Technical Indicator: Weak Uptrend.'
-            elif macd[-1] > macdsignal[-1] and macd[-1] > 0.0:
-                # 強い上昇トレンド
-                s.technical_indicator = DEF.TECHNICAL_UPTREND
-                msg = 'Technical Indicator: Strong Uptrend.'
-            elif macd[-1] < macdsignal[-1] and macd[-1] > 0.0:
-                # 弱い降下トレンド
-                s.technical_indicator = DEF.TECHNICAL_DOWNTREND
-                msg = 'Technical Indicator: Weak Downtrand.'
-            elif macd[-1] < macdsignal[-1] and macd[-1] < 0.0:
-                # 強い降下トレンド
-                s.technical_indicator = DEF.TECHNICAL_DOWNTREND
-                msg = 'Technical Indicator: Strong Downtrand.'
+            elif macd[-1] > macdsignal[-1]:
+                if macd[-1] < 0.0:
+                    # 弱い上昇トレンド
+                    s.technical_indicator = DEF.TECHNICAL_UPTREND
+                    msg = 'Technical Indicator: Weak Uptrend.'
+                else:
+                    # 強い上昇トレンド
+                    s.technical_indicator = DEF.TECHNICAL_UPTREND
+                    msg = 'Technical Indicator: Strong Uptrend.'
+            elif macd[-1] < macdsignal[-1]:
+                if macd[-1] > 0.0:
+                    # 弱い降下トレンド
+                    s.technical_indicator = DEF.TECHNICAL_DOWNTREND
+                    msg = 'Technical Indicator: Weak Downtrand.'
+                else:
+                    # 強い降下トレンド
+                    s.technical_indicator = DEF.TECHNICAL_DOWNTREND
+                    msg = 'Technical Indicator: Strong Downtrand.'
             else:
                 # いずれにも該当しない
                 s.technical_indicator = DEF.TECHNICAL_UNCHANGED