R - Finance back test function for entry / exit signal - r

Question

Ok so I am back testing trading signals using R. Here is a snippet of my code which shows the z-score creation, close to close returns (using TTR package), the long signal and the Lag() to place the signal the next day.
require(quantmod)
require(TTR)
require(zoo)
# Calculate n period close price z-scores indicator using TTR package
new.df$roll.mean.n3 <- runMean(new.df$Close, n=3, cumulative = FALSE)
new.df$roll.sd.n3 <- runSD(new.df$Close, n=3, cumulative = FALSE)
new.df$roll.z.score.n3 <- apply(new.df[,c('Close','roll.mean.n3', 'roll.sd.n3')], 1, function(x) { (x[1]-x[2])/x[3] } )
# Calculate Close-to-Close returns
new.df$clret <- ROC(new.df$Close,1)
new.df$clret[1] <- 0
# Create the long (up) signal
new.df$sigup <- ifelse(new.df$roll.z.score.n3 < -1, 1, 0)
# Lag signals to align with days in market not days when signals were generated
new.df$sigup <- Lag(new.df$sigup,1) # Note k=1 implies a move *forward*
The current setup above produces an output like this :
roll.z.score.n3 sigup
1 NA 0
2 NA 0
3 -1.135974424 0
4 0.193311168 1
5 0.714285714 0
6 -1.148753543 0
7 -0.942160394 1
8 -0.695763683 0
9 1.140646864 0
10 0.985196899 0
11 -0.768766574 0
12 -1.011293858 0
13 -0.516703612 1
14 -1.120897077 0
15 1.091089451 1
The entry signal is to go long when zscore value is <-1 which is shown in row 3. We have a +1 on row 4 because we used Lag() to forward step the entry signal to the next day. Each time the z-score value is below -1, there is a +1 the next day.
This setup is perfectly fine if i'm only trading for 1 holding day only.
I can then multiply sigup 1 x % daily returns to obtain an equity curve.
I want to elaborate further on the entry / exit signals. I wish to go long (sig long) when zscore is <-1 and exit when z-score is >1.
The output would look something like this:
roll.z.score.n3 sig long
1 NA 0
2 NA 0
3 -1.135974424 0
4 0.193311168 1
5 0.714285714 1
6 -1.148753543 1
7 -0.942160394 1
8 -0.695763683 1
9 1.140646864 1
10 0.985196899 0
11 -0.768766574 0
12 -1.011293858 0
13 -0.516703612 1
14 -1.120897077 1
15 1.091089451 1
16 0.968364052 0
17 0.872871561 0
18 1.099524999 0
19 0.918397948 0
Row 3 shows a zscore signal of <-1. Lag next day makes it +1 (row 4). And it stays +1 all way until row 9 when z-score signal is >1.0. Thus, the next day at row 10, the signal is 0.
I wanted to give some background on the current coding, its an attempt to further the post at FOSS trading blog.
Thanks for taking a look at this.

Answer 1

See if the following works:
zz = '
roll.z.score.n3 sig_long
1 NA 0
2 NA 0
3 -1.135974424 0
4 0.193311168 1
5 0.714285714 1
6 -1.148753543 1
7 -0.942160394 1
8 -0.695763683 1
9 1.140646864 1
10 0.985196899 0
11 -0.768766574 0
12 -1.011293858 0
13 -0.516703612 1
14 -1.120897077 1
15 1.091089451 1
16 0.968364052 0
17 0.872871561 0
18 1.099524999 0
19 0.918397948 0
'
df <- read.table(text = zz, header = TRUE)
df = na.omit(df)
df$sig_long[[1]] = ifelse(df$roll.z.score.n3[[1]] < (-1), 1, 0)
for (i in 2:nrow(df)){
df$sig_long[i] = ifelse(df$roll.z.score.n3[i] < (-1), 1,
ifelse(df$roll.z.score.n3[i] > 1, 0,
df$sig_long[i-1]))
}
Not sure about this part:
df$sig_long <- Lag(df$sig_long, 1)