options_backtester/asset_backtester/examples/example.ipynb at 3ec734508d0a86a5fb271af821b7b39b3c9432d2

In [1]:

import pandas as pd
import numpy as np
import os

In [2]:

from backtester import Backtest
from portfolio.portfolio import Portfolio
from portfolio.asset import Asset
from datahandler import HistoricalAssetData
from statistics.charts import returns_chart, returns_histogram, monthly_returns_heatmap

In [3]:

%config InlineBackend.figure_format="retina"
%matplotlib inline

In [4]:

import pandas_datareader.data as web
import datetime

As a first example, we run a backtest of a portfolio with the following tickers during the year 2019. Data is taken from yahoo finance.

In [5]:

start = datetime.datetime(2019, 1, 1)
end = datetime.datetime(2019, 12, 31)

tickers = ["VOO", "TUR", "RSX", "EWY", "EWS", "VTIP", "TLT", "BWX", "PDBC", "IAU", "VNQI"]

dfs = []
for ticker in tickers:
    df = web.DataReader(ticker, 'yahoo', start, end).reset_index().rename(columns={'Date': 'date', 'High': 'high', 'Low': 'low', 'Open': 'open', 'Close': 'close', 'Volume': 'volume'})
    df['symbol'] = ticker
    dfs.append(df)

In [9]:

df = pd.concat(dfs)

In [10]:

df = df.sort_values(by='date').reset_index(drop=True)

In [11]:

df.to_csv("data/portfolio_data.csv", index=False)

Use HistoricalAssetData to load your csv. Data must include 'date', 'Adj Close' and 'symbol' columns to work.

In [12]:

data = HistoricalAssetData("data/portfolio_data.csv")
schema = data.schema

In [13]:

data

Out [13]:

	date	high	low	open	close	volume	Adj Close	symbol
0	2019-01-02	230.850006	226.020004	226.179993	229.990005	4891300.0	225.350449	VOO
1	2019-01-02	15.333000	14.890000	14.920000	15.120000	1382500.0	14.909319	PDBC
2	2019-01-02	122.160004	121.339996	121.660004	122.150002	19841500.0	119.344078	TLT
3	2019-01-02	47.910000	47.860001	47.860001	47.910000	771400.0	46.986149	VTIP
4	2019-01-02	22.100000	21.850000	21.930000	22.100000	876200.0	21.076557	EWS
...	...	...	...	...	...	...	...	...
2767	2019-12-31	27.180000	27.000000	27.110001	27.090000	132900.0	27.090000	TUR
2768	2019-12-31	16.650000	16.510000	16.520000	16.559999	1780800.0	16.559999	PDBC
2769	2019-12-31	295.989990	294.170013	294.529999	295.799988	2512700.0	295.799988	VOO
2770	2019-12-31	136.460007	135.380005	136.210007	135.479996	10707400.0	135.479996	TLT
2771	2019-12-31	59.090000	58.849998	58.950001	59.090000	219900.0	59.090000	VNQI

2772 rows × 8 columns

To create a portfolio, use the Portfolio class and then create assets with Asset(name, percentage), where name should match the name given in the 'symbol' column and percentage is the percentage (from 0 to 1) allocated to that specific asset.

In [14]:

portfolio = Portfolio()

In [15]:

VOO = Asset('VOO', 0.1)
TUR = Asset('TUR', 0.05)
RSX = Asset('RSX', 0.05)
EWY = Asset('EWY', 0.05)
EWS = Asset('EWS', 0.05)
VTIP = Asset('VTIP', 0.10)
TLT = Asset('TLT', 0.20)
BWX = Asset('BWX', 0.10)
PDBC = Asset('PDBC', 0.05)
IAU = Asset('IAU', 0.15)
VNQI = Asset('VNQI', 0.10)

In [16]:

portfolio.add_assets([
    VOO, TUR, RSX, EWY, EWS, VTIP, TLT, BWX,
    PDBC, IAU, VNQI
])

Out [16]:

<portfolio.portfolio.Portfolio at 0x1165ef750>

Create a Backtest object passing it the schema of your data, add the portfolio and data to it and then run it. The run method takes the initial capital (default value 1.000.000) and periods as arguments. periods defines how often, in months, a rebalancing of the portfolio is made, so that a value of '1' means a monthly rebalancing and '6' a bi-annual one. Note that its value should be a string and it defaults to '1'. For a backtest with no rebalancing use None.

In [17]:

bt = Backtest(schema)
bt.portfolio = portfolio
bt.data = data

In [18]:

bt.run(initial_capital=1_000_000, periods=None)

Out [18]:

0% [██████████████████████████████] 100% | ETA: 00:00:00
Total time elapsed: 00:00:04

	capital	cash	total_value	% change	accumulated return
2019-01-01	1.000000e+06	1000000.000000	NaN	NaN	NaN
2019-01-02	1.000000e+06	403.112655	9.995969e+05	0.000000	1.000000
2019-01-03	9.987008e+05	403.112655	9.982976e+05	-0.001299	0.998701
2019-01-04	1.008706e+06	403.112655	1.008302e+06	0.010018	1.008706
2019-01-07	1.010929e+06	403.112655	1.010526e+06	0.002205	1.010929
...	...	...	...	...	...
2019-12-24	1.158018e+06	403.112655	1.157615e+06	0.002707	1.158018
2019-12-26	1.163238e+06	403.112655	1.162835e+06	0.004508	1.163238
2019-12-27	1.165936e+06	403.112655	1.165533e+06	0.002319	1.165936
2019-12-30	1.163417e+06	403.112655	1.163014e+06	-0.002160	1.163417
2019-12-31	1.163667e+06	403.112655	1.163264e+06	0.000215	1.163667

253 rows × 5 columns

When done, the backtester returns a balance sheet with the daily returns and wealth values. Pass this balance dataframe to the functions returns_chart, returns_histogram and monthly_returns_heatmap for better visualization.

In [19]:

returns_chart(bt.balance)

Out [19]:

In [20]:

returns_histogram(bt.balance)

Out [20]:

In [21]:

monthly_returns_heatmap(bt.balance)

Out [21]:

We now run a backtest of a portfolio from the book "The Ivy Portfolio" by Faber and Richardson, consisting of an even allocation in domestic (US) stocks (VTI), foreign stocks (VEU), bonds (BND), real estate (VNQ) and commodities (DBC).

The data we'll use consists of the last 10 years (2010-2019), once again taken from yahoo finance.

In [22]:

tickers =  ["VTI", "VEU", "BND", "VNQ", "DBC"]

dfs = []
for ticker in tickers:
    start = datetime.datetime(2020, 1, 1)
    end = datetime.datetime(2020, 12, 31)
    for i in range(10):
        start = start.replace(year = start.year - 1)
        end = end.replace(year = end.year - 1)
        df = web.DataReader(ticker, 'yahoo', start, end).reset_index().rename(columns={'Date': 'date', 'High': 'high', 'Low': 'low', 'Open': 'open', 'Close': 'close', 'Volume': 'volume'})
        df['symbol'] = ticker
        dfs.append(df)

In [23]:

ivy_df = pd.concat(dfs)

In [24]:

ivy_df = ivy_df.sort_values(by='date').reset_index(drop=True)

In [25]:

ivy_df.to_csv("data/ivy_portfolio_data.csv", index=False)

In [ ]:

In [26]:

data = HistoricalAssetData("data/ivy_portfolio_data.csv")
schema = data.schema

In [27]:

data

Out [27]:

	date	high	low	open	close	volume	Adj Close	symbol
0	2010-01-04	57.380001	56.840000	56.860001	57.310001	2251500.0	47.202740	VTI
1	2010-01-04	78.730003	78.540001	78.599998	78.680000	1098100.0	58.797794	BND
2	2010-01-04	44.959999	44.400002	44.660000	44.959999	1692500.0	33.440495	VEU
3	2010-01-04	45.419998	44.200001	45.220001	44.549999	2408400.0	29.869335	VNQ
4	2010-01-04	25.240000	25.070000	25.170000	25.240000	2046900.0	24.523520	DBC
...	...	...	...	...	...	...	...	...
12575	2019-12-31	16.049999	15.910000	15.950000	15.950000	806100.0	15.950000	DBC
12576	2019-12-31	53.790001	53.480000	53.619999	53.750000	1688400.0	53.750000	VEU
12577	2019-12-31	92.830002	92.000000	92.089996	92.790001	4649500.0	92.790001	VNQ
12578	2019-12-31	83.919998	83.769997	83.849998	83.860001	5931300.0	83.860001	BND
12579	2019-12-31	163.759995	162.720001	163.000000	163.619995	3262400.0	163.619995	VTI

12580 rows × 8 columns

In [28]:

portfolio = Portfolio()

In [29]:

VTI = Asset("VTI", 0.2)
VEU = Asset("VEU", 0.2)
BND = Asset("BND", 0.2)
VNQ = Asset("VNQ", 0.2)
DBC = Asset("DBC", 0.2)

In [30]:

portfolio.add_assets([VTI, VEU, BND, VNQ, DBC])

Out [30]:

<portfolio.portfolio.Portfolio at 0x116500490>

In [31]:

bt = Backtest(schema)
bt.portfolio = portfolio
bt.data = data

In [32]:

# No rebalancing
bt.run(initial_capital = 1_000_000, periods=None)

Out [32]:

0% [██████████████████████████████] 100% | ETA: 00:00:00
Total time elapsed: 00:00:19

	capital	cash	total_value	% change	accumulated return
2010-01-03	1.000000e+06	1000000.000000	NaN	NaN	NaN
2010-01-04	1.000000e+06	92.035288	9.999080e+05	0.000000	1.000000
2010-01-05	1.001321e+06	92.035288	1.001229e+06	0.001321	1.001321
2010-01-06	1.005620e+06	92.035288	1.005528e+06	0.004293	1.005620
2010-01-07	1.004723e+06	92.035288	1.004631e+06	-0.000892	1.004723
...	...	...	...	...	...
2019-12-24	2.039190e+06	92.035288	2.039098e+06	0.001160	2.039190
2019-12-26	2.048080e+06	92.035288	2.047988e+06	0.004360	2.048080
2019-12-27	2.050816e+06	92.035288	2.050724e+06	0.001336	2.050816
2019-12-30	2.045149e+06	92.035288	2.045056e+06	-0.002764	2.045149
2019-12-31	2.051284e+06	92.035288	2.051192e+06	0.003000	2.051284

2517 rows × 5 columns

In [33]:

returns_chart(bt.balance)

Out [33]:

In [34]:

monthly_returns_heatmap(bt.balance)

Out [34]:

For comparison, you can run the same backtest on Portfolio Visualizer.

In [35]:

# Monthly rebalancing
bt.run(initial_capital = 1_000_000, periods='1')

Out [35]:

0% [██████████████████████████████] 100% | ETA: 00:00:00
Total time elapsed: 00:00:22

	capital	cash	total_value	% change	accumulated return
2010-01-03	1.000000e+06	1000000.000000	NaN	NaN	NaN
2010-01-04	1.000000e+06	92.035288	9.999080e+05	0.000000	1.000000
2010-01-05	1.001321e+06	92.035288	1.001229e+06	0.001321	1.001321
2010-01-06	1.005620e+06	92.035288	1.005528e+06	0.004293	1.005620
2010-01-07	1.004723e+06	92.035288	1.004631e+06	-0.000892	1.004723
...	...	...	...	...	...
2019-12-24	1.817790e+06	57.640553	1.817733e+06	0.001500	1.817790
2019-12-26	1.826130e+06	57.640553	1.826072e+06	0.004588	1.826130
2019-12-27	1.829028e+06	57.640553	1.828970e+06	0.001587	1.829028
2019-12-30	1.824079e+06	57.640553	1.824021e+06	-0.002706	1.824079
2019-12-31	1.826055e+06	57.640553	1.825997e+06	0.001083	1.826055

2517 rows × 5 columns

In [36]:

returns_chart(bt.balance)

Out [36]:

In [ ]:

In [37]:

# Bi-annual rebalancing
bt.run(initial_capital = 1_000_000, periods='6')

Out [37]:

0% [██████████████████████████████] 100% | ETA: 00:00:00
Total time elapsed: 00:00:21

	capital	cash	total_value	% change	accumulated return
2010-01-03	1.000000e+06	1000000.000000	NaN	NaN	NaN
2010-01-04	1.000000e+06	92.035288	9.999080e+05	0.000000	1.000000
2010-01-05	1.001321e+06	92.035288	1.001229e+06	0.001321	1.001321
2010-01-06	1.005620e+06	92.035288	1.005528e+06	0.004293	1.005620
2010-01-07	1.004723e+06	92.035288	1.004631e+06	-0.000892	1.004723
...	...	...	...	...	...
2019-12-24	1.823260e+06	131.209041	1.823129e+06	0.001476	1.823260
2019-12-26	1.831628e+06	131.209041	1.831497e+06	0.004590	1.831628
2019-12-27	1.834523e+06	131.209041	1.834392e+06	0.001580	1.834523
2019-12-30	1.829496e+06	131.209041	1.829364e+06	-0.002740	1.829496
2019-12-31	1.831590e+06	131.209041	1.831458e+06	0.001145	1.831590

2517 rows × 5 columns

In [38]:

returns_chart(bt.balance)

Out [38]:

In [39]:

monthly_returns_heatmap(bt.balance)

Out [39]:

In [40]:

returns_histogram(bt.balance)

Out [40]:

In [ ]:

3.1 MiB Raw Blame History Unescape Escape

3.1 MiB

Raw Blame History