Home > Backtesting, R, Trading Strategies > Introduction to Backtesting library in the Systematic Investor Toolbox

Introduction to Backtesting library in the Systematic Investor Toolbox

I wrote a simple Backtesting library to evaluate and analyze Trading Strategies. I will use this library to present the performance of trading strategies that I will study in the next series of posts.

It is very easy to write a simple Backtesting routine in R, for example:

bt.simple <- function(data, signal) 
	# lag singal
	signal = Lag(signal, 1)

	# back fill
	signal = na.locf(signal, na.rm = FALSE)
	signal[is.na(signal)] = 0

	# calculate Close-to-Close returns
	ret = ROC(Cl(data), type='discrete')
	ret[1] = 0
	# compute stats	
	bt = list()
		bt$ret = ret * signal
		bt$equity = cumprod(1 + bt$ret)    	    	

# Test for bt.simple functions
# load historical prices from Yahoo Finance
data = getSymbols('SPY', src = 'yahoo', from = '1980-01-01', auto.assign = F)

# Buy & Hold
signal = rep(1, nrow(data))
buy.hold = bt.simple(data, signal)
# MA Cross
sma = SMA(Cl(data),200)
signal = ifelse(Cl(data) > sma, 1, 0)
sma.cross = bt.simple(data, signal)
# Create a chart showing the strategies perfromance in 2000:2009
dates = '2000::2009'
buy.hold.equity = buy.hold$equity[dates] / as.double(buy.hold$equity[dates][1])
sma.cross.equity = sma.cross$equity[dates] / as.double(sma.cross$equity[dates][1])

chartSeries(buy.hold.equity, TA = c(addTA(sma.cross.equity, on=1, col='red')),	
	theme ='white', yrange = range(buy.hold.equity, sma.cross.equity) )	

The code I implemented in the Systematic Investor Toolbox is a bit longer, but follows the same logic. It provides extra functionality: ability to handle multiple securities, weights or shares backtesting, and customized reporting. Following is a sample code to implement the above strategies using the backtesting library in the Systematic Investor Toolbox:

# Load Systematic Investor Toolbox (SIT)
con = gzcon(url('https://github.com/systematicinvestor/SIT/raw/master/sit.gz', 'rb'))

	# Load historical data
	tickers = spl('SPY')

	data <- new.env()
	getSymbols(tickers, src = 'yahoo', from = '1970-01-01', env = data, auto.assign = T)
	bt.prep(data, align='keep.all', dates='1970::2011')

	# Code Strategies
	prices = data$prices    
	# Buy & Hold	
	data$weight[] = 1
	buy.hold = bt.run(data)	

	# MA Cross
	sma = bt.apply(data, function(x) { SMA(Cl(x), 200) } )	
	data$weight[] = NA
		data$weight[] = iif(prices >= sma, 1, 0)
	sma.cross = bt.run(data, trade.summary=T)			

	# Create Report
	plotbt.custom.report(sma.cross, buy.hold)

The bt.prep function merges and aligns all symbols in the data environment. The bt.apply function applies user given function to each symbol in the data environment. The bt.run computes the equity curve of strategy specified by data$weight matrix. The data$weight matrix holds weights (signals) to open/close positions. The plotbt.custom.report function creates the customized report, which can be fined tuned by the user. Here is a sample output:

> buy.hold = bt.run(data)
Performance summary :
        CAGR    Best    Worst
        7.2     14.5    -9.9

> sma.cross = bt.run(data, trade.summary=T)
Performance summary :
        CAGR    Best    Worst
        6.3     5.8     -7.2

The visual performance summary:

The statistical performance summary:

The trade summary:

To view the complete source code for this example, please have a look at the bt.test() function in bt.r at github.

  1. November 25, 2011 at 8:50 pm

    very nicely done. Like the report best of all. Why did you choose to limit the drawdown and ROC to models[1]? Do you mind if I use in a post to demonstrate how to change the weights, or also to potentially tie in with ttrTests?

  2. November 25, 2011 at 10:06 pm

    Kent, thank you and please go ahead and use the Systematic Investor Toolbox in your posts.

    I’m writing a series of posts that will explain functionality and present examples of the Backtesting library.

    I decided to include only the first model ( models[1] ) in the “12 Month Rolling” and “Drawdown” plots to limit the clutter. You can create your own version of plotbt.custom.report function to customize reporting. For example, plotbt.timelyportfolio.report function.

  1. No trackbacks yet.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: