require(quantmod)
getSymbols("GLD")
for(i in seq(5, 15, 5))
for(j in seq(50, 80, 15))
{
GLD$fast <- SMA(Cl(GLD), n=i)
GLD$slow <- SMA(Cl(GLD), n=j)
golden_cross <- Lag(ifelse(GLD$fast > GLD$slow, 1, -1))
golden_cross <- na.locf(golden_cross, na.rm=TRUE)
coin <- ROC(Cl(GLD))*golden_cross
best_coin <- max(coin)
worst_coin <- min(coin)
last_coin <- cumprod(1+coin)[NROW(coin),]
annual_coin <- round((last_coin-1)*100, digits=2)/(NROW(coin)/252)
cat(i,j,annual_coin, best_coin, worst_coin, "\n", file="~/goldcat", append=TRUE)
cat(i,j,annual_coin, best_coin, worst_coin, "\n")
}
Milktrader
Milktrader
Saturday, April 9, 2011
A Super-Easy, Simple-Dimple Backtester in R
I cut my finger on a paring knife this morning. Don't use a sharp knife to spread butter on your toast. It's better to limit yourself to using dull kitchen utensils until the caffeine kicks in. No matter, I still have most of my digits to type in a simple backtesting program in R. Good thing it's only about 15 lines of code.
Subscribe to:
Post Comments (Atom)
I think you missed a few lines there.....
ReplyDeleteThe majority of the respectable paid computer registry optimizers that exist on the Internet are good and are also not too costly.
Delete@Andreas fixed it. Was missing the closing bracket at the very end.
ReplyDeletethanks Milky I am enjoying your posts
ReplyDeleteAt first it seems not easy..
ReplyDeleteHi. Appreciate your blog! I'm very new to R and in fact my mind has been tarnished by years programming in C style static languages so I'm having a little problem getting to grips with the magic of R. A few questions that I came across while looking at this code, would be awesome if you could provide some hints:
ReplyDelete1)ifelse(GLD$fast > GLD$slow, 1, -1)--> This does just a comparison of the SMAs correct? I'd be more interested to find out the actual crossover of the SMA i.e.
GLD$fast[n-1]GLD$slow[n]
2)I've seen the documentation to check what lag and na.locf do but they break my code when I remove the NAs with this error
"Error: dims [product 23989] do not match the length of object [23994]"
while running coin <- ROC(eurusd$close)*golden_cross
3) Due to the above I've skipped of the few NAs to the code written below - but I'm not 100% sure if it works :). Also the coin_mean should be a good indicator, correct?
eurusd$fast <- SMA(eurusd$close, n=i)
eurusd$slow <- SMA(eurusd$close, n=j)
cross <- ifelse(eurusd$fast > eurusd$slow, 1, -1)#cross <- Lag(ifelse(eurusd$fast > eurusd$slow, 1, -1))
#cross <- na.locf(cross, na.rm=TRUE)
coin <- ROC(eurusd$close)*cross
#coin <- na.locf(coin, na.rm=TRUE)
best_coin <- max(na.locf(coin, na.rm=TRUE))#best_coin <- max(coin)
worst_coin <- min(na.locf(coin, na.rm=TRUE))#worst_coin <- min(coin)
last_coin <- cumprod(1+coin)[NROW(coin)] # last_coin <- cumprod(1+coin)[NROW(coin),]
mean_coin <- mean(na.locf(coin, na.rm=TRUE))
cat(i,j, best_coin, worst_coin, mean_coin, "\n", file=op, append=TRUE)
cat(i,j, best_coin, worst_coin, mean_coin, "\n")
I've actually gone ahead and implemented a cross over(at least I think I have :) style the way I discussed above:
ReplyDeletefor(i in seq(2, 5))
{
for(j in seq(6, 15))
{
eurusd$fast<-SMA(eurusd$close, n=i)
eurusd$slow<-SMA(eurusd$close,n=j)
cross <- NA
for(n in 2:length(eurusd$close))
{
cross[n] <- ifelse(eurusd$fast[n] > eurusd$slow[n] && eurusd$fast[n-1] < eurusd$slow[n-1] , 1, (ifelse(eurusd$fast[n] < eurusd$slow[n] && eurusd$fast[n-1] > eurusd$slow[n-1], -1, 0)))
}
I'm guessing there might be a way to improve performance but this is all I could conjure up with the little R experience I have..
I love what you created for helping people do R backtesters. I have created a platform to backtest in various languages including R. I've also included a large list of built in indicators.
ReplyDeletecheck it out, R backtester documentation
cheers,
Joshua
Quantonomics
I wish you wouldn't spam these forums with
Deleteyour (yet) non existent website.
Error in xts(new.x, x.index) :
ReplyDeleteorder.by requires an appropriate time-based object
I get this error when running the code?