'ES_base_price_change', 'ES_test_start_price', 'ES_test_end_price',
'ES_test_price_change', 'ES_test_price_abs_change', 'RTY_base_start_price',
'RTY_base_end_price', 'RTY_base_price_change', 'RTY_test_start_price',
'RTY_test_end_price', 'RTY_test_price_change', 'RTY_test_price_abs_change'
]]
for test_trade_day in trade_day_list:
test_trade_day_list = [test_trade_day]
max_profit = -10000
for base_time_end in base_time_end_list:
test_time_start = base_time_end
for test_length in test_length_list:
test_time_end = (ts.cal_datetime(test_time_start) +
datetime.timedelta(minutes=test_length)).time()
ES_base_timeframe_df = ts.select_date_timeframe(
ES_Tele.df, test_trade_day, base_time_start, base_time_end)
ES_test_timeframe_df = ts.select_date_timeframe(
ES_Tele.df, test_trade_day, test_time_start, test_time_end)
RTY_base_timeframe_df = ts.select_date_timeframe(
RTY_Tele.df, test_trade_day, base_time_start, base_time_end)
RTY_test_timeframe_df = ts.select_date_timeframe(
RTY_Tele.df, test_trade_day, test_time_start, test_time_end)
try:
ES_base_start_price = ES_base_timeframe_df['TradePrice'][0]
ES_base_end_price = ES_base_timeframe_df['TradePrice'][-1]
ES_base_price_change = ES_base_end_price - ES_base_start_price
ES_test_start_price = ES_test_timeframe_df['TradePrice'][0]
ES_test_end_price = ES_test_timeframe_df['TradePrice'][-1]
ES_test_price_change = ES_test_end_price - ES_test_start_price
ES_test_price_abs_change = np.absolute(ES_test_price_change)
def sma_theodiff_strategy(Tele2, trade_day_list, base_time_start, base_time_end, test_time_start, test_time_end, sma_threshold, stop_threshold, bounce_stop_threshold, over_limit_length, trade_cost):
import numpy as np
import pandas as pd
import telescope as ts
import telescope_metrics as tm
import datetime
from scipy.stats.stats import pearsonr
from tqdm import tqdm
#--------------Input Parameters---------------
# data_mpk_file_path, start_date, end_date
# df_2019 = data_mpk_file_path
# START = start_date
# END = end_date
# STIME = base_time_start
# ETIME = test_time_end
# TIME_FRAME_LENGTH = time_frame_length
# ROLLING_STEP = rolling_step #(has to be a factor of TIME_FRAME_LENGTH)
BASE_TIME_START = base_time_start
BASE_TIME_END = base_time_end
TEST_TIME_START = test_time_start
TEST_TIME_END = test_time_end
SMA_THRESHOLD = sma_threshold
STOP_THRESHOLD = stop_threshold
TRADE_COST = trade_cost
#---------------------------------------------
#--------------Input Parameters---------------
#df_2019 = r"Z:\KaiData\Theo\2019\YM.ESTheo.mpk"
#START = datetime.datetime(2019, 7, 8, 0, 0)
#END = datetime.datetime(2019, 7, 12, 23, 59)
#STIME = datetime.time(8, 0, 0)
#ETIME = datetime.time(15, 15, 0)
#TIME_FRAME_LENGTH = 5
#ROLLING_STEP = '1T' #(has to be a factor of TIME_FRAME_LENGTH)
#---------------------------------------------
#print(Tele2.df)
#print(Tele2.grouped)
#print(Tele2.num_of_groups)
#print(Tele2.group_names)
# time_list = Tele2.df['time'].unique()
daily_profit_list = []
limit_stop_count = 0
bounce_stop_count = 0
trade_count = 0
# check_list = []
loss_trade_day_info = [['Date', 'Enter SMA Ratio', 'Entered Position', 'Exit Time', 'Exit Reason', 'Buy Price', 'Sell Price', 'Net Profit']]
for trade_day in trade_day_list:
base_timeframe_df = ts.select_date_timeframe(Tele2.df, trade_day, BASE_TIME_START, BASE_TIME_END)
test_timeframe_df = ts.select_date_timeframe(Tele2.df, trade_day, TEST_TIME_START, TEST_TIME_END)
profit = 0
base_bool_series = (base_timeframe_df['TheoPrice'] > base_timeframe_df['SMA'])
above_ratio = base_bool_series.sum()/base_bool_series.size
below_ratio = 1 - above_ratio
bounce_stop_reach_count = 0
if (np.absolute(base_timeframe_df['TheoPrice'][-1] - base_timeframe_df['TheoPrice'][0])) >=100:
if above_ratio >= SMA_THRESHOLD:
enter_ratio = above_ratio
enter_position = 'Long'
exit_reason = 'Exit Time Reach'
try:
buy_price = test_timeframe_df[test_timeframe_df['time'] == test_time_start]['TheoPrice'][0]
except:
print(test_timeframe_df['time'])
print(test_time_start)
max_price = buy_price
for time, price in test_timeframe_df['TheoPrice'].items():
# if price <= buy_price - STOP_THRESHOLD:
# sell_price = price
# close_time = time
# limit_stop_count += 1
# exit_reason = 'Reach Loss Up Stop'
# break
if price > max_price:
max_price = price
bounce_stop_reach_count = 0
if price < max_price - bounce_stop_threshold:
bounce_stop_reach_count += 1
if bounce_stop_reach_count >= over_limit_length:
sell_price = price
close_time = time
bounce_stop_count += 1
exit_reason = 'Reach Bounce Stop'
break
sell_price = price
close_time = time
# check_list.append([sell_price, buy_price, close_time])
trade_count += 1
profit = sell_price - buy_price - TRADE_COST
if below_ratio > SMA_THRESHOLD:
enter_ratio = below_ratio
enter_position = 'Short'
exit_reason = 'Exit Time Reach'
sell_price = test_timeframe_df[test_timeframe_df['time'] == test_time_start]['TheoPrice'][0]
min_price = sell_price
for time, price in test_timeframe_df['TheoPrice'].items():
# if price >= sell_price + STOP_THRESHOLD:
# buy_price = price
# close_time = time
# limit_stop_count +=1
# exit_reason = 'Reach Loss Up Stop'
# break
if price < min_price:
min_price = price
bounce_stop_reach_count = 0
if price > min_price + bounce_stop_threshold:
bounce_stop_reach_count += 1
if bounce_stop_reach_count >= over_limit_length:
buy_price = price
close_time = time
bounce_stop_count += 1
exit_reason = 'Reach Bounce Stop'
break
buy_price = price
close_time = time
trade_count += 1
profit = sell_price - buy_price - TRADE_COST
# check_list.append([sell_price, buy_price, close_time])
daily_profit_list.append(profit)
if (profit < 0):
day_info = [trade_day, enter_ratio, enter_position, close_time.time(), exit_reason, buy_price, sell_price, profit]
loss_trade_day_info.append(day_info)
time_stop_count = trade_count - limit_stop_count - bounce_stop_count
return np.array(daily_profit_list), trade_count, limit_stop_count, bounce_stop_count, time_stop_count, loss_trade_day_info
# result = pd.DataFrame(column_list[1:], columns=column_list[0])
# result['time'] = time_list[:-ignore_col_num]
# result.set_index('time', inplace=True)
# result.to_excel(output_excel_path)
# return result
def consecutive_5_backtest(Tele2, trade_day_list, base_time_start,
base_time_end, test_time_start, test_time_end,
sma_threshold, stop_threshold,
bounce_stop_threshold, trade_cost):
import numpy as np
import pandas as pd
import telescope as ts
import telescope_metrics as tm
import datetime
from scipy.stats.stats import pearsonr
from tqdm import tqdm
#--------------Input Parameters---------------
# data_mpk_file_path, start_date, end_date
# df_2019 = data_mpk_file_path
# START = start_date
# END = end_date
# STIME = base_time_start
# ETIME = test_time_end
# TIME_FRAME_LENGTH = time_frame_length
# ROLLING_STEP = rolling_step #(has to be a factor of TIME_FRAME_LENGTH)
BASE_TIME_START = base_time_start
BASE_TIME_END = base_time_end
TEST_TIME_START = test_time_start
TEST_TIME_END = test_time_end
SMA_THRESHOLD = sma_threshold
STOP_THRESHOLD = stop_threshold
TRADE_COST = trade_cost
#---------------------------------------------
#--------------Input Parameters---------------
#df_2019 = r"Z:\KaiData\Theo\2019\YM.ESTheo.mpk"
#START = datetime.datetime(2019, 7, 8, 0, 0)
#END = datetime.datetime(2019, 7, 12, 23, 59)
#STIME = datetime.time(8, 0, 0)
#ETIME = datetime.time(15, 15, 0)
#TIME_FRAME_LENGTH = 5
#ROLLING_STEP = '1T' #(has to be a factor of TIME_FRAME_LENGTH)
#---------------------------------------------
#print(Tele2.df)
#print(Tele2.grouped)
#print(Tele2.num_of_groups)
#print(Tele2.group_names)
# time_list = Tele2.df['time'].unique()
daily_profit_list = []
limit_stop_count = 0
bounce_stop_count = 0
trade_count = 0
# check_list = []
for trade_day in trade_day_list:
base_timeframe_df = ts.select_date_timeframe(Tele2.df, trade_day,
BASE_TIME_START,
BASE_TIME_END)
test_timeframe_df = ts.select_date_timeframe(Tele2.df, trade_day,
TEST_TIME_START,
TEST_TIME_END)
profit = 0
# base_bool_series = (base_timeframe_df['TheoPrice'] > base_timeframe_df['SMA'])
# above_ratio = base_bool_series.sum()/base_bool_series.size
# below_ratio = 1 - above_ratio
# if above_ratio >= SMA_THRESHOLD:
# try:
# buy_price = test_timeframe_df[test_timeframe_df['time'] == test_time_start]['TheoPrice'][0]
# except:
# print(test_timeframe_df['time'])
# print(test_time_start)
# max_price = buy_price
# for time, price in test_timeframe_df['TheoPrice'].items():
# if price <= buy_price - STOP_THRESHOLD:
# sell_price = price
# limit_stop_count += 1
# break
# if price > max_price:
# max_price = price
# if price < max_price - bounce_stop_threshold:
# sell_price = price
# bounce_stop_count += 1
# break
# sell_price = price
## close_time = time
#
## check_list.append([sell_price, buy_price, close_time])
# trade_count += 1
# profit = sell_price - buy_price - TRADE_COST
#
# if below_ratio > SMA_THRESHOLD:
# sell_price = test_timeframe_df[test_timeframe_df['time'] == test_time_start]['TheoPrice'][0]
# min_price = sell_price
# for time, price in test_timeframe_df['TheoPrice'].items():
# if price >= sell_price + STOP_THRESHOLD:
# buy_price = price
# limit_stop_count +=1
# break
# if price < min_price:
# min_price = price
# if price > min_price + bounce_stop_threshold:
# buy_price = price
# bounce_stop_count += 1
# break
# buy_price = price
## close_time = time
# trade_count += 1
# profit = sell_price - buy_price - TRADE_COST
# check_list.append([sell_price, buy_price, close_time])
daily_profit_list.append(profit)
time_stop_count = trade_count - limit_stop_count - bounce_stop_count
return np.array(
daily_profit_list
), trade_count, limit_stop_count, bounce_stop_count, time_stop_count
# result = pd.DataFrame(column_list[1:], columns=column_list[0])
# result['time'] = time_list[:-ignore_col_num]
# result.set_index('time', inplace=True)
# result.to_excel(output_excel_path)
# return result