def SMA_cross(data):
sma50 = data.rolling(50).mean()
sma200 = data.rolling(200).mean()
tw = sma200.copy()
tw[sma50 > sma200] = 1.0
tw[sma50 <= sma200] = -1.0
tw[sma200.isnull()] = 0.0
tmp = bt.merge(tw, data, sma50, sma200)
tmp.columns = ["tw", "price", "sma50", "sma200"]
# tmp.columns = ['tw', 'price', 'sma50', 'sma200']
ax = tmp.plot(figsize = (15, 5), secondary_y = ["tw"])
plt.savefig("smacross.png")
ma_cross = bt.Strategy("ma_cross",
[WeighTarget(tw),
bt.algos.Rebalance()
])
t = bt.Backtest(ma_cross, data)
res = bt.run(t)
res.display()
res.plot()
plt.savefig("smacross_res")
def SMA(data):
sma = data.rolling(50).mean()
plot = bt.merge(data, sma).plot(figsize=(15, 5))
plt.savefig("sma.png")
# 建立策略
s = bt.Strategy("above50sma",
[SelectWhere(data > sma),
bt.algos.WeighEqually(),
bt.algos.Rebalance()
])
# 建立回测
t = bt.Backtest(s, data)
# 执行回测
res = bt.run(t)
# 输出结果
res.display()
res.plot()
plt.savefig("sma_result.png")
# 多种策略的测试
sma10 = above_sma(data, sma_per = 10, name = "sma10")
sma20 = above_sma(data, sma_per = 10, name = "sma20")
sma40 = above_sma(data, sma_per = 10, name = "sma40")
base_line = baseTest(data)
# 一起运行回测
res2 = bt.run(sma10, sma20, sma40, base_line)
# 输出结果
res2.display()
res2.plot()
plt.savefig("sma.png")
plt.savefig("multi_sma_result.png")
import bt
import talib
# Construct the signal
signal[stock_rsi > 70] = -1
signal[stock_rsi < 30] = 1
signal[(stock_rsi <= 70) & (stock_rsi >= 30)] = 0
# Merge the data
combined_df = bt.merge(signal, price_data)
combined_df.columns = ['signal', 'Price']
combined_df.plot(secondary_y=['signal'])
plt.show()
# Define the strategy
bt_strategy = bt.Strategy('RSI_MeanReversion',
[bt.algos.WeighTarget(signal),
bt.algos.Rebalance()])
# Create the backtest and run it
bt_backtest = bt.Backtest(bt_strategy, price_data)
bt_result = bt.run(bt_backtest)
# Plot the backtest result
bt_result.plot(title='Backtest result')
plt.show()
### Finds optimal SMA cross settings and return percentage # [best_low, best_high, best_percentage] = sma_cross_optimal_values(data, sma_low_bot, sma_low_top, sma_high_bot, sma_high_top, within_percentage) ### Get SMA based on the data # sma_low_plot = pd.rolling_mean(data, best_low) # sma_high_plot = pd.rolling_mean(data, best_high) # # plot = bt.merge(data, sma_low_plot, sma_high_plot).plot() ############################ ### Finds optimal SMA value for price crossing the SMA line # [best_sma, best_percentage] = above_sma_optimal_value(data, 5, sma_high_top, within_percentage) ### Get SMA based on the data # sma_best_plot = pd.rolling_mean(data, best_sma) # Do final test to print full results test = above_sma(data, best_sma) results = bt.run(test) # Display graphical plot of the data plot = bt.merge(data, sma_best_plot).plot() results.plot() results.display_monthly_returns() results.display() # Necessary to show matlibplot graphic produced with res.plot() plt.show() results.values()
#all_loaded_data = load_data('600519,000858,601318,600036,603288,600276,600900,600887', do_normalize_data)
all_loaded_data = load_data('600732', do_normalize_data)
bench_data = load_data('sh000001', do_normalize_data)
start_date = '20210501'
end_date = '20210630'
print(start_date, end_date)
all_data = panel = create_dataframe(all_loaded_data, 'close')
bench_data = create_dataframe(bench_data, 'close')
panel = filter_dataframe(panel, start_date, end_date)
bench_data = filter_dataframe(bench_data, start_date, end_date)
bt.merge(panel).plot()
plt.show()
# create the strategy
from strategy.rsi_25_75_talib import create_strategy
from strategy.r3 import create_strategy as r3_create_strategy
from strategy.sma import above_sma, long_only_ew
from strategy.macd_talib import create_strategy as macd_create_strategy
from strategy.ibs import create_strategy as ibs_create_strategy
from strategy.ibs_rsi import create_strategy as ibs_rsi_create_strategy
ss = bt.Strategy('s1', [
bt.algos.RunMonthly(),
bt.algos.SelectAll(),
bt.algos.WeighEqually(),
bt.algos.Rebalance()
import bt
import matplotlib.pyplot as plt
# Construct the signal
signal[EMA_short > EMA_long] = 1
signal[EMA_short < EMA_long] = -1
# Merge the data
combined_df = bt.merge(signal, price_data, EMA_short, EMA_long)
combined_df.columns = ['Signal', 'Price', 'EMA_short', 'EMA_long']
# Plot the signal, price and MAs
combined_df.plot(secondary_y=['Signal'])
plt.show()
# Define the strategy
bt_strategy = bt.Strategy('EMA_crossover',
[bt.algos.WeighTarget(signal),
bt.algos.Rebalance()])
# Create the backtest and run it
bt_backtest = bt.Backtest(bt_strategy, price_data)
bt_result = bt.run(bt_backtest)
# Plot the backtest result
bt_result.plot(title='Backtest result')
plt.show()
def rsi_strat(ticker: str, start_date: Union[datetime, str],
other_args: List[str]):
"""Strategy that buys when the stock is less than a threshold and shorts when it exceeds a threshold.
Parameters
----------
ticker : str
Stock to test
start : Union[str, datetime]
Backtest start date. Can be either string or datetime
other_args : List[str]
List of argparse arguments
"""
parser = argparse.ArgumentParser(
add_help=False,
formatter_class=argparse.ArgumentDefaultsHelpFormatter,
prog="rsi_strat",
description="""Strategy that buys when the stock is less than a threshold
and shorts when it exceeds a threshold.""",
)
parser.add_argument(
"-p",
"--periods",
dest="periods",
help="Number of periods for RSI calculation",
type=check_positive,
default=14,
)
parser.add_argument(
"-u",
"--high",
default=70,
dest="high",
type=check_positive,
help="High (upper) RSI Level",
)
parser.add_argument("-l",
"--low",
default=30,
dest="low",
type=check_positive,
help="Low RSI Level")
parser.add_argument(
"--spy",
action="store_true",
default=False,
help="Flag to add spy hold comparison",
dest="spy",
)
parser.add_argument(
"--no_bench",
action="store_true",
default=False,
help="Flag to not show buy and hold comparison",
dest="no_bench",
)
parser.add_argument(
"--no_short",
action="store_false",
default=True,
dest="shortable",
help="Flag that disables the short sell",
)
try:
ns_parser = parse_known_args_and_warn(parser, other_args)
if not ns_parser:
return
if ns_parser.high < ns_parser.low:
print("Low RSI value is higher than Low RSI value\n")
return
ticker = ticker.lower()
# prices = bt.get(ticker, start=start_date)
prices = get_data(ticker, start_date)
rsi = pd.DataFrame(ta.rsi(prices[ticker], ns_parser.periods))
rsi.columns = [ticker]
signal = 0 * rsi.copy()
signal[rsi > ns_parser.high] = -1
signal[rsi < ns_parser.low] = 1
signal[rsi.isnull()] = 0
merged_data = bt.merge(signal, prices)
merged_data.columns = ["signal", "price"]
bt_strategy = bt.Strategy(
"RSI Reversion",
[bt.algos.WeighTarget(signal),
bt.algos.Rebalance()])
bt_backtest = bt.Backtest(bt_strategy, prices)
if ns_parser.spy:
spy_bt = buy_and_hold("spy", start_date, "SPY Hold")
if ns_parser.no_bench:
res = bt.run(bt_backtest, spy_bt)
else:
stock_bt = buy_and_hold(ticker, start_date,
ticker.upper() + " Hold")
res = bt.run(bt_backtest, spy_bt, stock_bt)
else:
if ns_parser.no_bench:
res = bt.run(bt_backtest)
else:
stock_bt = buy_and_hold(ticker, start_date,
ticker.upper() + " Hold")
res = bt.run(bt_backtest, stock_bt)
plot_bt(res,
f"RSI Strategy between ({ns_parser.low}, {ns_parser.high})")
print(res.display(), "\n")
except Exception as e:
print(e, "\n")
def ema_cross(ticker: str, start_date: Union[str, datetime],
other_args: List[str]):
"""Strategy where we go long/short when EMA(short) is greater than/less than EMA(short)
Parameters
----------
ticker : str
Stock to test
start : Union[str, datetime]
Backtest start date. Can be either string or datetime
other_args : List[str]
List of argparse arguments
"""
parser = argparse.ArgumentParser(
add_help=False,
formatter_class=argparse.ArgumentDefaultsHelpFormatter,
prog="ema_cross",
description=
"Cross between a long and a short Exponential Moving Average.",
)
parser.add_argument(
"-l",
"--long",
default=50,
dest="long",
type=check_positive,
help="Long EMA period",
)
parser.add_argument(
"-s",
"--short",
default=20,
dest="short",
type=check_positive,
help="Short EMA period",
)
parser.add_argument(
"--spy",
action="store_true",
default=False,
help="Flag to add spy hold comparison",
dest="spy",
)
parser.add_argument(
"--no_bench",
action="store_true",
default=False,
help="Flag to not show buy and hold comparison",
dest="no_bench",
)
parser.add_argument(
"--no_short",
action="store_false",
default=True,
dest="shortable",
help="Flag that disables the short sell",
)
try:
ns_parser = parse_known_args_and_warn(parser, other_args)
if not ns_parser:
return
if ns_parser.long < ns_parser.short:
print("Short EMA period is longer than Long EMA period\n")
return
ticker = ticker.lower()
# prices = bt.get(ticker, start=start_date)
prices = get_data(ticker, start_date)
short_ema = pd.DataFrame(ta.ema(prices[ticker], ns_parser.short))
short_ema.columns = [ticker]
long_ema = pd.DataFrame(ta.ema(prices[ticker], ns_parser.long))
long_ema.columns = [ticker]
# signals
signals = long_ema.copy()
signals[short_ema > long_ema] = 1.0
signals[short_ema <= long_ema] = -1.0 * ns_parser.shortable
signals[long_ema.isnull()] = 0.0
combined_data = bt.merge(signals, prices, short_ema, long_ema)
combined_data.columns = ["signal", "price", "ema_short", "ema_long"]
bt_strategy = bt.Strategy(
"EMA_Cross",
[
bt.algos.WeighTarget(signals),
bt.algos.Rebalance(),
],
)
bt_backtest = bt.Backtest(bt_strategy, prices)
if ns_parser.spy:
spy_bt = buy_and_hold("spy", start_date, "SPY Hold")
if ns_parser.no_bench:
res = bt.run(bt_backtest, spy_bt)
else:
stock_bt = buy_and_hold(ticker, start_date,
ticker.upper() + " Hold")
res = bt.run(bt_backtest, spy_bt, stock_bt)
else:
if ns_parser.no_bench:
res = bt.run(bt_backtest)
else:
stock_bt = buy_and_hold(ticker, start_date,
ticker.upper() + " Hold")
res = bt.run(bt_backtest, stock_bt)
plot_bt(res,
f"EMA Cross for EMA({ns_parser.short})/EMA({ns_parser.long})")
print(res.display(), "\n")
except Exception as e:
print(e, "\n")
import bt
import matplotlib
matplotlib.get_backend()
# download data
from bt.algos import SelectWhere
data = bt.get('aapl,msft', start='2016-01-01')
# calculate moving average DataFrame using pandas' rolling_mean
import pandas as pd
# a rolling mean is a moving average, right?
sma = pd.rolling_mean(data, 50)
bt.merge(data, sma).plot(figsize=(15, 5))
signal = data > sma
# first we create the Strategy
s = bt.Strategy('above50sma', [SelectWhere(data > sma),
bt.algos.WeighEqually(),
bt.algos.Rebalance()])
# now we create the Backtest
t = bt.Backtest(s, data)
# and let's run it!
res = bt.run(t)
res.plot('d')
print(type(res))
res.display()
name,
[bt.algos.WeighTarget(tw), bt.algos.Rebalance()])
# Set up the backtest
uvxy_dynamic = bt.Backtest(strategy_4_uvxy,
strategy_4_data[['UVXY']],
integer_positions=False)
# Why integer positions? See UVXY head data
print(strategy_4_data['UVXY'].head(10))
# Run the backtest
res = bt.run(uvxy_dynamic, benchmark_test)
# Once we ran the previous backtest, we can extract the prices data to create synthetic securities
synthetic = bt.merge(res['uvxy_dynamic'].prices, res['long_spy'].prices)
# This is our new data, which is essentially the equity curve of each sub-strategy we tested.
# We can use this data to test our final strategy, just as before.
strategy_4 = bt.Strategy('combined_uvxy_spy', [
bt.algos.SelectAll(),
bt.algos.WeighSpecified(uvxy_dynamic=0.5, long_spy=0.5),
bt.algos.Rebalance()
])
# Create and run
t = bt.Backtest(strategy_4, synthetic, integer_positions=False)
res = bt.run(t, benchmark_test)
# Display summary statistics and plot the weights
res.display()
if __name__ == "__main__":
# 获取数据
# data = getData()
#data = bt.get("spy", start = "2010-01-01")
#print(data.head())
#print(data.describe())
# SMA(data)
#SMA_cross(data)
t1 = sma_cross("aapl", name = "aapl_mass_cross")
t2 = sma_cross("msft", name = "msft_mass_cross")
res = bt.run(t1, t2)
data = bt.merge(res["aapl_mass_cross"].prices,
res["msft_mass_cross"].prices)
s = bt.Strategy('s',
[bt.algos.SelectAll(),
bt.algos.WeighInvVol(),
bt.algos.Rebalance()
])
t = bt.Backtest(s, data)
res = bt.run(t)
res.display()
res.plot()
plt.savefig("multi_smacross.png")
res.plot_weights()
plt.savefig("multi_smacross_weights.png")
def rsi_strategy(
ticker: str,
df_stock: pd.DataFrame,
periods: int,
low_rsi: int,
high_rsi: int,
spy_bt: bool = True,
no_bench: bool = False,
shortable: bool = True,
) -> bt.backtest.Result:
"""Perform backtest for simple EMA strategy. Buys when price>EMA(l)
Parameters
----------
ticker : str
Stock ticker
df_stock : pd.DataFrame
Dataframe of prices
periods : int
Number of periods for RSI calculati
low_rsi : int
Low RSI value to buy
hirh_rsi : int
High RSI value to sell
spy_bt : bool
Boolean to add spy comparison
no_bench : bool
Boolean to not show buy and hold comparison
shortable : bool
Flag to disable the ability to short sell
Returns
-------
Result
Backtest results
"""
ticker = ticker.lower()
start_date = df_stock.index[0]
prices = pd.DataFrame(df_stock["Adj Close"])
prices.columns = [ticker]
rsi = pd.DataFrame(ta.rsi(prices[ticker], periods))
rsi.columns = [ticker]
signal = 0 * rsi.copy()
signal[rsi > high_rsi] = -1 * shortable
signal[rsi < low_rsi] = 1
signal[rsi.isnull()] = 0
merged_data = bt.merge(signal, prices)
merged_data.columns = ["signal", "price"]
bt_strategy = bt.Strategy(
"RSI Reversion", [bt.algos.WeighTarget(signal),
bt.algos.Rebalance()])
bt_backtest = bt.Backtest(bt_strategy, prices)
bt_backtest = bt.Backtest(bt_strategy, prices)
backtests = [bt_backtest]
if spy_bt:
spy_bt = buy_and_hold("spy", start_date, "SPY Hold")
backtests.append(spy_bt)
if not no_bench:
stock_bt = buy_and_hold(ticker, start_date, ticker.upper() + " Hold")
backtests.append(stock_bt)
res = bt.run(*backtests)
return res
def ema_cross_strategy(
ticker: str,
df_stock: pd.DataFrame,
short_length: int,
long_length: int,
spy_bt: bool = True,
no_bench: bool = False,
shortable: bool = True,
) -> bt.backtest.Result:
"""Perform backtest for simple EMA strategy. Buys when price>EMA(l)
Parameters
----------
ticker : str
Stock ticker
df_stock : pd.DataFrame
Dataframe of prices
short_length : int
Length of short ema window
long_length : int
Length of long ema window
spy_bt : bool
Boolean to add spy comparison
no_bench : bool
Boolean to not show buy and hold comparison
shortable : bool
Boolean to allow for selling of the stock at cross
Returns
-------
Result
Backtest results
"""
ticker = ticker.lower()
start_date = df_stock.index[0]
prices = pd.DataFrame(df_stock["Adj Close"])
prices.columns = [ticker]
short_ema = pd.DataFrame(ta.ema(prices[ticker], short_length))
short_ema.columns = [ticker]
long_ema = pd.DataFrame(ta.ema(prices[ticker], long_length))
long_ema.columns = [ticker]
# signals
signals = long_ema.copy()
signals[short_ema > long_ema] = 1.0
signals[short_ema <= long_ema] = -1.0 * shortable
signals[long_ema.isnull()] = 0.0
combined_data = bt.merge(signals, prices, short_ema, long_ema)
combined_data.columns = ["signal", "price", "ema_short", "ema_long"]
bt_strategy = bt.Strategy(
"EMA_Cross",
[
bt.algos.WeighTarget(signals),
bt.algos.Rebalance(),
],
)
bt_backtest = bt.Backtest(bt_strategy, prices)
backtests = [bt_backtest]
if spy_bt:
spy_bt = buy_and_hold("spy", start_date, "SPY Hold")
backtests.append(spy_bt)
if not no_bench:
stock_bt = buy_and_hold(ticker, start_date, ticker.upper() + " Hold")
backtests.append(stock_bt)
res = bt.run(*backtests)
return res
import bt
import matplotlib
matplotlib.get_backend()
# download data
from bt.algos import SelectWhere
data = bt.get('aapl,msft', start='2016-01-01')
# calculate moving average DataFrame using pandas' rolling_mean
import pandas as pd
# a rolling mean is a moving average, right?
sma = pd.rolling_mean(data, 50)
bt.merge(data, sma).plot(figsize=(15, 5))
signal = data > sma
# first we create the Strategy
s = bt.Strategy(
'above50sma',
[SelectWhere(data > sma),
bt.algos.WeighEqually(),
bt.algos.Rebalance()])
# now we create the Backtest
t = bt.Backtest(s, data)
# and let's run it!
res = bt.run(t)
res.plot('d')
# save in temp - this will be used by the weighing algo
target.temp['selected'] = selected
# return True because we want to keep on moving down the stack
return True
data = bt.get('aapl,msft,c,gs,ge', start='2010-01-01')
# calculate moving average DataFrame using pandas' rolling_mean
import pandas as pd
# a rolling mean is a moving average, right?
sma = pd.rolling_mean(data, 50)
# let's see what the data looks like - this is by no means a pretty chart, but it does the job
tmp = bt.merge(data, sma).asfreq('m', 'ffill').rebase()
plot = tmp.plot(figsize=(15, 5))
signal = data > sma
# first we create the Strategy
s = bt.Strategy('above50sma', [SelectWhere(data > sma),
bt.algos.WeighEqually(),
bt.algos.Rebalance()])
# now we create the Backtest
t = bt.Backtest(s, data)
# and let's run it!
res = bt2.run(t)