Improved Moving Average?

I have been notified by the authors that the code does not perfectly reflect the improved moving average introduced.  In another post, I will explore the differences.  The authors have now released their version of the code http://www.quantf.com/fotis-papailias/improved-moving-average-code-is-available-for-download/332.

When @quantfblog started following me on Twitter, I was delighted to discover their papers

Papailias, Fotis and Thomakos, Dimitrios D., An Improved Moving Average Technical Trading Rule (September 11, 2011). Available at SSRN: http://ssrn.com/abstract=1926376

Papailias, Fotis and Thomakos, Dimitrios D., An Improved Moving Average Technical Trading Rule II - Can We Obtain Performance Improvements with Short Sales? (November 12, 2011). Available at SSRN: http://ssrn.com/abstract=1958906

backed by a nice and improving website http://www.quantf.com.  I just could not resist the opportunity to port their improved moving average idea to R and run some additional tests.  The entire process was extremely pleasant due to the authors’ willingness to test, comment, and suggest throughout the implementation process.  Thanks so much to them for all their help.

In this process, I AM NOT OFFERING INVESTMENT ADVICE.  THIS IS SIMPLY A TEST/ILLUSTRATION.  PURSUING THESE CONCEPTS WILL MOST LIKELY LOSE SIGNIFICANT AMOUNTS (ALL) OF YOUR MONEY.

 

From TimelyPortfolio

From TimelyPortfolio

From TimelyPortfolio

For fun, I thought it would be interesting to compare the "Improved Moving Average" to a Mebane Faber style 10-month moving average system.

From TimelyPortfolio

 

While I enjoyed the testing, I am still not entirely sure if the “improved moving average” is significantly improved, but it certainly might fit someone’s utility curve better than the standard moving average.  More than anything, this process has proven to me a couple of things:

1) the beauty of open-source and collaboration.  The authors were incredibly generous and helpful as I worked through this process.  To even better demonstrate the power of open-source, I will use ttrTests to do additional testing and then the bt examples from http://systematicinvestor.wordpress.com in future posts.

2) how even a simple moving average can become incredibly complex.  Making one very slight change markedly changes the results.

R code from GIST:

 

	#get MAImp code from GIST
	#thanks to http://systematicinvestor.wordpress.com
	#for showing me how to do this
	con=url("https://raw.github.com/gist/1405187/92e7c24ff2459a0830c45b828f1dba41143e9436/MAImp.r")
	source(con)
	require(quantmod)
	require(PerformanceAnalytics)
	getSymbols("^GSPC",from = "1896-01-01", to = Sys.Date())
	#test 20 day simple crossover 50 day simple
	signals <- MAImp(GSPC, k1=20, k2=50)
	#another example would be a simple 100 day moving average price cross
	#signals <- MAImp(GSPC, k1=100)
	#use lag 1 to eliminate foresight
	#signal is 1 (in) and 0 (out)
	ret <- merge(rep(ROC(GSPC[,4],type="discrete",n=1),3) * lag(signals,k=1),ROC(GSPC[,4],type="discrete",n=1))
	colnames(ret) <- c(colnames(signals),"BuyHold")
	#jpeg(filename="performance summary.jpg",
	quality=100,width=6.25, height = 8, units="in",res=96)
	charts.PerformanceSummary(ret,ylog=TRUE,
	main = "Test of Improved Moving Average",
	colorset=c("cadetblue","darkolivegreen3","gray70","bisque3"))
	#dev.off()
	#table.AnnualizedReturns(ret)
	require(ggplot2)
	#ggplot downside statistics
	downside <- as.data.frame(table.DownsideRisk(ret))
	#get rid of drawdown since it messes up the chart
	#and is already shown on Performance Summary
	downside <- downside[c(1:6,8:11),]
	downsideMelt<-melt(cbind(rownames(downside),downside),id.vars=1)
	colnames(downsideMelt)<-c("Statistic","System","Value")
	#jpeg(filename="ggplot downside.jpg",
	quality=100,width=6.25, height = 8, units="in",res=96)
	ggplot(downsideMelt, stat="identity", aes(x=Statistic,y=Value,fill=System)) +
	geom_bar(position="dodge") +
	coord_flip() +
	scale_fill_manual(values=c("cadetblue","darkolivegreen3","gray70","bisque3","purple")) +
	theme_bw()
	#dev.off()
	#ggplot CAPM statistics
	capm <- as.data.frame(table.CAPM(ret[,1:3],ret[,4]))
	capmTable<-melt(cbind(rownames(capm),capm),id.vars=1)
	colnames(capmTable)<-c("Statistic","System","Return")
	#jpeg(filename="ggplot capm.jpg",
	quality=100,width=6.25, height = 8, units="in",res=96)
	ggplot(capmTable, stat="identity", aes(x=Statistic,y=Return,fill=System)) +
	geom_bar(position="dodge") +
	coord_flip() +
	scale_fill_manual(values=c("cadetblue","darkolivegreen3","gray70","bisque3","purple")) +
	theme_bw()
	#dev.off()
	#test 10 month Mebane Faber style moving average for another example
	GSPC.monthly <- to.monthly(GSPC)[,4]
	#convert date index to yyyy-mm-dd
	index(GSPC.monthly) <- as.Date(index(GSPC.monthly))
	signals.monthly <- MAImp(GSPC.monthly, k1=10)
	#use lag 1 to eliminate foresight
	#signal is 1 (in) and 0 (out)
	ret.monthly <- merge(rep(ROC(GSPC.monthly,type="discrete",n=1),3) * lag(signals.monthly,k=1),ROC(GSPC.monthly,type="discrete",n=1))
	colnames(ret.monthly) <- c(colnames(signals.monthly),"BuyHold")
	#jpeg(filename="performance monthly.jpg",
	quality=100,width=6.25, height = 8, units="in",res=96)
	charts.PerformanceSummary(ret.monthly,ylog=TRUE,
	main = "Test of Improved Moving Average on Monthly",
	colorset=c("cadetblue","darkolivegreen3","gray70","bisque3"))
	#dev.off()

view raw MAImp.test.r hosted with ❤ by GitHub