-
Notifications
You must be signed in to change notification settings - Fork 131
/
backtest.py
202 lines (155 loc) · 7.21 KB
/
backtest.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
#-------------------------------------------------------------------------------
# Name: backtest
# Purpose: perform routine backtesting tasks.
# This module should be useable as a stand-alone library outide of the TWP package.
#
# Author: Jev Kuznetsov
#
# Created: 03/07/2014
# Copyright: (c) Jev Kuznetsov 2013
# Licence: BSD
#-------------------------------------------------------------------------------
import pandas as pd
import matplotlib.pyplot as plt
import sys
import numpy as np
def tradeBracket(price,entryBar,upper=None, lower=None, timeout=None):
'''
trade a bracket on price series, return price delta and exit bar #
Input
------
price : numpy array of price values
entryBar: entry bar number, *determines entry price*
upper : high stop
lower : low stop
timeout : max number of periods to hold
Returns exit price and number of bars held
'''
assert isinstance(price, np.ndarray) , 'price must be a numpy array'
# create list of exit indices and add max trade duration. Exits are relative to entry bar
if timeout: # set trade length to timeout or series length
exits = [min(timeout,len(price)-entryBar-1)]
else:
exits = [len(price)-entryBar-1]
p = price[entryBar:entryBar+exits[0]+1] # subseries of price
# extend exits list with conditional exits
# check upper bracket
if upper:
assert upper>p[0] , 'Upper bracket must be higher than entry price '
idx = np.where(p>upper)[0] # find where price is higher than the upper bracket
if idx.any():
exits.append(idx[0]) # append first occurence
# same for lower bracket
if lower:
assert lower<p[0] , 'Lower bracket must be lower than entry price '
idx = np.where(p<lower)[0]
if idx.any():
exits.append(idx[0])
exitBar = min(exits) # choose first exit
return p[exitBar], exitBar
class Backtest(object):
"""
Backtest class, simple vectorized one. Works with pandas objects.
"""
def __init__(self,price, signal, signalType='capital',initialCash = 0, roundShares=True):
"""
Arguments:
*price* Series with instrument price.
*signal* Series with capital to invest (long+,short-) or number of shares.
*sitnalType* capital to bet or number of shares 'capital' mode is default.
*initialCash* starting cash.
*roundShares* round off number of shares to integers
"""
#TODO: add auto rebalancing
# check for correct input
assert signalType in ['capital','shares'], "Wrong signal type provided, must be 'capital' or 'shares'"
#save internal settings to a dict
self.settings = {'signalType':signalType}
# first thing to do is to clean up the signal, removing nans and duplicate entries or exits
self.signal = signal.ffill().fillna(0)
# now find dates with a trade
tradeIdx = self.signal.diff().fillna(0) !=0 # days with trades are set to True
if signalType == 'shares':
self.trades = self.signal[tradeIdx] # selected rows where tradeDir changes value. trades are in Shares
elif signalType =='capital':
self.trades = (self.signal[tradeIdx]/price[tradeIdx])
if roundShares:
self.trades = self.trades.round()
# now create internal data structure
self.data = pd.DataFrame(index=price.index , columns = ['price','shares','value','cash','pnl'])
self.data['price'] = price
self.data['shares'] = self.trades.reindex(self.data.index).ffill().fillna(0)
self.data['value'] = self.data['shares'] * self.data['price']
delta = self.data['shares'].diff() # shares bought sold
self.data['cash'] = (-delta*self.data['price']).fillna(0).cumsum()+initialCash
self.data['pnl'] = self.data['cash']+self.data['value']-initialCash
@property
def sharpe(self):
''' return annualized sharpe ratio of the pnl '''
pnl = (self.data['pnl'].diff()).shift(-1)[self.data['shares']!=0] # use only days with position.
return sharpe(pnl) # need the diff here as sharpe works on daily returns.
@property
def pnl(self):
'''easy access to pnl data column '''
return self.data['pnl']
def plotTrades(self):
"""
visualise trades on the price chart
long entry : green triangle up
short entry : red triangle down
exit : black circle
"""
l = ['price']
p = self.data['price']
p.plot(style='x-')
# ---plot markers
# this works, but I rather prefer colored markers for each day of position rather than entry-exit signals
# indices = {'g^': self.trades[self.trades > 0].index ,
# 'ko':self.trades[self.trades == 0].index,
# 'rv':self.trades[self.trades < 0].index}
#
#
# for style, idx in indices.iteritems():
# if len(idx) > 0:
# p[idx].plot(style=style)
# --- plot trades
#colored line for long positions
idx = (self.data['shares'] > 0) | (self.data['shares'] > 0).shift(1)
if idx.any():
p[idx].plot(style='go')
l.append('long')
#colored line for short positions
idx = (self.data['shares'] < 0) | (self.data['shares'] < 0).shift(1)
if idx.any():
p[idx].plot(style='ro')
l.append('short')
plt.xlim([p.index[0],p.index[-1]]) # show full axis
plt.legend(l,loc='best')
plt.title('trades')
class ProgressBar:
def __init__(self, iterations):
self.iterations = iterations
self.prog_bar = '[]'
self.fill_char = '*'
self.width = 50
self.__update_amount(0)
def animate(self, iteration):
print '\r',self,
sys.stdout.flush()
self.update_iteration(iteration + 1)
def update_iteration(self, elapsed_iter):
self.__update_amount((elapsed_iter / float(self.iterations)) * 100.0)
self.prog_bar += ' %d of %s complete' % (elapsed_iter, self.iterations)
def __update_amount(self, new_amount):
percent_done = int(round((new_amount / 100.0) * 100.0))
all_full = self.width - 2
num_hashes = int(round((percent_done / 100.0) * all_full))
self.prog_bar = '[' + self.fill_char * num_hashes + ' ' * (all_full - num_hashes) + ']'
pct_place = (len(self.prog_bar) // 2) - len(str(percent_done))
pct_string = '%d%%' % percent_done
self.prog_bar = self.prog_bar[0:pct_place] + \
(pct_string + self.prog_bar[pct_place + len(pct_string):])
def __str__(self):
return str(self.prog_bar)
def sharpe(pnl):
return np.sqrt(250)*pnl.mean()/pnl.std()