R - please share your experiences - page 5

 
The problem with R is that it cannot be run in conjunction with a tester and graphical reports cannot be drawn. =)
 
wise:
The problem with R is that it cannot be run in conjunction with a tester and graphical reports cannot be drawn. =)
What prevents it?
 
I don't know. I just saw people on forexfactory talking about it. Maybe I got it wrong. =)
 
Hmm. Maybe I don't know either. Haven't really used it yet.
 
wise:
The problem with R is that you can't run it in conjunction with the tester and draw graphical reports. =)


This is not a problem of R, it's a problem of report drawers.

For research it is enough to write a simple tester by yourself in five minutes.

library(TTR)

library(zoo)

# get some data

num.points <- ...

midprices <- ...

spreads <- ...

bids <- midprices - 0.5 * spreads

asks <- midprices + 0.5 * spreads

# length(midprices) == num.points; length(spreads) == num.points

ma.fast <- EMA(midprices, 15)

ma.slow <- EMA(midprices, 50)

# compute entry/exit signals

open.long.position <- (ma.fast > ma.slow) & (ma.slow > midprices)

close.long.position <- (ma.fast < midprices)

open.short.position <- (ma.fast < ma.slow) & (ma.slow < midprices)

close.short.position <- (ma.fast > midprices)

signals.to.position <- function (num.points, signal.open, signal.close) {

pos <- rep(NA, num.points)

pos[signal.open] <- 1

pos[signal.close] <- 0

pos <- na.locf(pos, na.rm = F)

pos[is.na(pos)] <- 0

pos

}

# compute strategy positions

pos.long <- signals.to.position(num.points, open.long.position, close.long.position)

pos.short <- signals.to.position(num.points, open.short.position, close.short.position)

pos.total <- pos.long - pos.short

# compute equity

commission <- 1

trade.size <- c(0, pos.total[2 : num.points] - pos.total[1 : (num.points - 1)])

fees <- abs(trade.size) * (commission + 0.5 * spreads)

equity <- midprices * pos.total + cumsum(-midprices * trade.size - fees)

# compute balance

balance <- rep(NA, num.points)

balance[trade.size != 0] <- equity[trade.size != 0]

balance <- na.locf(balance, na.rm = F)

balance[is.na(balance)] <- 0

# display results

par(mfrow = c(2, 1))

plot(midprices, t = 'l', col = 'gray', lty = 'dashed')

lines(bids, col = 'blue')

lines(asks, col = 'red')

points(which(trade.size > 0, arr.ind = T), asks[trade.size > 0], col = 'blue')

points(which(trade.size < 0, arr.ind = T), bids[trade.size < 0], col = 'red')

plot(equity, t = 'l')

lines(balance, col = 'gray', lty = 'dashed')

 
Has anyone been able to get MT5 + R to work?
 
anonymous:


That's not the problem of R, it's the problem of the report drawers.

All it takes is five minutes to write a simple tester yourself for research.

library(TTR)

library(zoo)

# get some data

num.points <- ...

midprices <- ...

spreads <- ...

bids <- midprices - 0.5 * spreads

asks <- midprices + 0.5 * spreads

# length(midprices) == num.points; length(spreads) == num.points

ma.fast <- EMA(midprices, 15)

ma.slow <- EMA(midprices, 50)

# compute entry/exit signals

open.long.position <- (ma.fast > ma.slow) & (ma.slow > midprices)

close.long.position <- (ma.fast < midprices)

open.short.position <- (ma.fast < ma.slow) & (ma.slow < midprices)

close.short.position <- (ma.fast > midprices)

signals.to.position <- function (num.points, signal.open, signal.close) {

pos <- rep(NA, num.points)

pos[signal.open] <- 1

pos[signal.close] <- 0

pos <- na.locf(pos, na.rm = F)

pos[is.na(pos)] <- 0

pos

}

# compute strategy positions

pos.long <- signals.to.position(num.points, open.long.position, close.long.position)

pos.short <- signals.to.position(num.points, open.short.position, close.short.position)

pos.total <- pos.long - pos.short

# compute equity

commission <- 1

trade.size <- c(0, pos.total[2 : num.points] - pos.total[1 : (num.points - 1)])

fees <- abs(trade.size) * (commission + 0.5 * spreads)

equity <- midprices * pos.total + cumsum(-midprices * trade.size - fees)

# compute balance

balance <- rep(NA, num.points)

balance[trade.size != 0] <- equity[trade.size != 0]

balance <- na.locf(balance, na.rm = F)

balance[is.na(balance)] <- 0

# display results

par(mfrow = c(2, 1))

plot(midprices, t = 'l', col = 'gray', lty = 'dashed')

lines(bids, col = 'blue')

lines(asks, col = 'red')

points(which(trade.size > 0, arr.ind = T), asks[trade.size > 0], col = 'blue')

points(which(trade.size < 0, arr.ind = T), bids[trade.size < 0], col = 'red')

plot(equity, t = 'l')

lines(balance, col = 'gray', lty = 'dashed')

Simple as that...

Shall we decipher the grids?

 
TheXpert:
Hmm. Maybe I don't know either. Haven't really used it yet.
Maybe because of the asynchrony? R can work asynchronously and the tester can't. But asynchrony is still exotic.
 
tara:

It's just, like everything...

Shall we decipher the grids?

Eh.... would be R the terminal language, not mql.
 

Please help.

I have entered EURUSD in R. I have calculated the model and calculated the coefficient. How can I draw a chart and combine it with the kotir?

> x.ar<-ar(eur[1:256],method="mle")

> x.ar


Call:

ar(x = eur[1:256],method="mle")


Coefficients:

1 ........2.......... 3

0.9420 0.1955 -0.1644


Order selected 3 sigma^2 estimated as 2.73e-06

Reason: