Friday, October 7, 2011

ttrTests 4th and Final Test

Hopefully you have been able to persist with me through

ttrTests This is a Test Test 3:Data Snoopy

ttrTests This is a Test--Test 1 and Test 2

ttrTests: Its Great Thesis and Incredible Potentia...

For the 4th and final ttrTest, we will check for the persistence of parameters across subperiods and then across subperiods with bootstrapped samples for each subperiod.  Not surprisingly, CUD failed most of the tests.  However, reader iQuant is one step ahead of me in his comment:

“Nice research piece, thank you! I was wondering whether we should consider the interest earned in cash positions as well the cost of going short. In long term strategies such as the one presented by Professor St. John this may make a lot of difference, favouring low frequency parameter sets and increasing the conditional excess returns of zero-weigth positions.

October 2, 2011 12:46 PM

I will attempt to address this very valid concern in future posts.

From TimelyPortfolio
From TimelyPortfolio

For the most intense test, we will bootstrap multiple samples across multiple subperiods. The prettiness of the graph unfortunately does not indicate the power of the test.

From TimelyPortfolio

R code (click to download from Google Docs):

#let's define our silly countupdown function
#as a sample of a custom ttr rule
CUD <- function(x,params=50,...) {
#CUD takes the n-period sum of 1 (up days) and -1 (down days)
temp <- ifelse(runSum(ifelse(ROC(x,1,type="discrete") > 0,1,-1),params)>=0,1,0)
#replace NA with 0 at beginning of period
temp[] <- 0
}   require(ttrTests)
require(PerformanceAnalytics)   #defaults functions is overridden by ggplot2 and plyr if loaded
#and will cause problems if you want to use ttrTests concurrently   tckrs <- c("GSPC","RUT","N225","GDAXI","DJUBS")   #use 1 or GSPC but adjust however you would like
test_price <- as.vector(get(tckrs[i])[,4])   #run subperiods and paramPersist to test for luck across subperiods
#"asks whether or not good choices of parameters"
#"were robust across different time periods"
#chose 6 since data is from 1950 will approximate by decade
subper <- subperiods(x=test_price, periods = 6, ttr = CUD,
start = 20, nSteps = 30, stepSize = 10, restrict = FALSE,
burn = 0, short = FALSE, condition = NULL,
silent = TRUE, TC = 0.001, loud = TRUE, alpha = 0.025,
file = "", benchmark = "hold")   #make output slightly more usable with some naming
#believe I got this right
names(subper[[2]]) <- "obs.correlation"
#while we are in a nasty for loop; grab some data also
for (j in 3:length(subper)) {
names(subper[[j]]) <- paste(c("excess.return","z.score","adj.excess.return",
# add this if desired ".subper",j-2, sep="")
ifelse(j==3, excess.df <- cbind(rep(j-2,length(subper[[j]]$tested.parameters)),
excess.df <- rbind(excess.df,
excess.df <-
colnames(excess.df) <- c("subperiod","parameter","excess.ret")   #run boxplot of excess returns by parameter
quality=100,width=6.25, height = 6.25, units="in",res=96)
xlab="Parameter", ylab="Excess Return",
main="Boxplot of Excess Returns by Parameter")   #jpeg(filename="strip chart.jpg",
# quality=100,width=6.25, height = 6.25, units="in",res=96)
stripchart(excess.df$excess.ret~excess.df$parameter, pch=19,
xlab="Parameter", ylab="Excess Return", vertical=TRUE,
main="Stripchart of Excess Returns by Parameter")     #and my favorite of all
#"tests if the persistence measure from subperiods()"
#"is statistically significant"
#this takes the longest (about 28 minutes on my i7 laptop)
#if you want to play
#change periods to 2 or bsamples to 10 to speed time
parpersist <- paramPersist(x=test_price, ttr = CUD, periods=6,
start = 20, nSteps = 30, stepSize = 10,
restrict = FALSE, burn = 0, short = FALSE, condition = NULL,
silent = TRUE, TC = 0.001, loud = TRUE, alpha = 0.025,
file = "")
names(parpersist) <- c("act.corr","obs.corr.samples","p.value")   #jpeg(filename="paramPersist correlations.jpg",
# quality=100,width=6.25, height = 6.25, units="in",res=96)
main="paramPersist for CUD")
text(0,parpersist$act.corr, "actual", col = "darkslateblue", adj = c(0, -.1))   #make output slightly more usable with some naming
#believe I got this right
names(snoop) <- c("details","V1","V2",
"V3","p1.for.l","p2.for.c","p3.for.u")   #jpeg(filename="dataSnoop values.jpg",
quality=100,width=6.25, height = 6.25, units="in",res=96)
type="l", col=2,
main="ttrTests dataSnoop V1,V2,and V3 on CUD",
xlab="Bootstrap Sample", ylab="Values")
points(snoop$V2, type="l", col=3)
points(snoop$V1, col=4)

Created by Pretty R at

1 comment:

  1. Hi,

    Thanks for a great series of posts. Learned a lot.

    I have some issues with the code:

    > names(snoop) <- c("details","V1","V2",
    + "V3","p1.for.l","p2.for.c","p3.for.u")
    Error in names(snoop) <- c("details", "V1", "V2", "V3", "p1.for.l", "p2.for.c", :
    object 'snoop' not found