[R-SIG-Finance] Specify time segment to calculate indicator and trade ?
Peter Chan
peter at returnandrisk.com
Thu Apr 9 09:12:40 CEST 2015
Hi there
There's a few ways to do this but the way that works for me (eg easy to debug) is to pre-process the data before letting quantstrat work its magic.
- create a custom indicator function that returns an xts object with same index as your price data
- add all your own indicators in columns eg if time between 10:00 and 14:00 --> Col1 = 1 otherwise 0
- once everything is correct and debugged you can use "add.indicator" to hook it up to quantstrat etc
I used this procedure in my blog on the FOMC cycle trading strategy.
If you have a lot of conditions, in the last column of your custom indicator you could multiply the appropriate binary (1, 0) cols together to get a "composite indicator" of when all indicators are true.
Then use that last column as your reference column in "add.signal" eg if ColComposite = 1, long entry
Thanks
Peter
---- On Thu, 09 Apr 2015 03:17:09 +0000 domodo<1111938 at qq.com> wrote ----
> hi,guys.
> quantstrat is very powerful and convient when coding strategy and backtest
> it with indicator and pre-define price pattern.but i'm not familiar with
> specify the strategy to trade in a segment of trading-time,below are two
> cases:
> 1)establish position from 10:00 to 14:00 only,and close position with
> technical exits
> 2)get the highest and lowest price of the first hour of a day
>
> for example,the strategy below is about price and single sma cross over.
> for case 1,i should store price series' date&time in varible 'datetime' as:
> [code]
> datetime <- as.POSIXlt(index(C))
> [/code]
>
> set entry time segment as:
> [code]
> entrytime1 <- 10
> entrytime2 <- 14
> [/code]
>
> and add some code below to judge whether the entry signal is in trading time
> :
> [code]
> signal <- ifelse(lag(C)>lag(MA) & lag(C,2)<lag(MA,2),1,
> ifelse(lag(C)<lag(MA) & lag(C,2)>lag(MA,2),-1,0))
>
> signal <- ifelse(signal>0 & datetime$hour>entrytime1 &
> datetime$hour<entrytime2,1,0)
> signal <- ifelse(signal<0 & datetime$hour>entrytime1 &
> datetime$hour<entrytime2,-1,0)
> [/code]
>
> but the code above is not so elegant at all,and low-efficient.
>
> for the case 2,i should code a function to calculate the first hour's
> highest & lowest price of each day,my idea in this function is similiar
> to code for case 1,that's, store price series's date&time as a POSIXlt
> variable 'datetime',and judge whether hour == 10,min == 15,sec==0,but this
> idea is also rough and low-efficient.
>
> any help or tips please ?
>
> below is the example strategy
>
> [code]
> library(blotter)
> library(quantstrat)
>
> #initialize environtments
> currency("USD")
> startdate <- '2012-01-18 09:14:00'
> finaldate <- '2012-01-19 15:14:00'
>
> future("if2", currency = "USD", multiplier = 300,
> tick_size = 0.2)
> Sys.setenv(TZ = "UTC")
> rm(list=ls(envir=.blotter),envir=.blotter)
> b.strategy <- "strategy"
>
> try(rm(list=ls(pos=.blotter),pos=.blotter),silent=TRUE)
> initPortf(b.strategy, "if2", initDate = startdate)
> initAcct(b.strategy, portfolios = b.strategy, initDate = startdate, initEq =
> 1000000)
>
> ifBAC2 <- read.table("C:/ifBAC2.csv", head = F, sep =
> ",")
> coredata <- ifBAC2[3:6]
> rownames(coredata) <- as.POSIXlt(paste(ifBAC2[,1],ifBAC2[,2]))
> ifxts <- as.xts(coredata)
> colnames(ifxts) <-
> c("open","high","low","close")
> if2 <- ifxts['2012-01-18 09:15:00/2012-01-19 15:14:00']
> if2$SMA15 <- SMA(Cl(if2),15)
>
> #custom theme
> myTheme <- chart_theme()
> myTheme$col$dn.col <- "lightgreen"
> myTheme$col$up.col <- "red"
> myTheme$col$dn.border <- "grey"
> myTheme$col$up.border <- "grey"
>
> MA <- if2$SMA15
> C <- Cl(if2)
> O <- Op(if2)
>
> #trading signal judgement
> signal <- ifelse(lag(C)>lag(MA) & lag(C,2)<lag(MA,2),1,
> ifelse(lag(C)<lag(MA) & lag(C,2)>lag(MA,2),-1,0))
> signal[is.na(signal)] <- 0
>
> #Bar-by-bar treatment
> for( i in 1:nrow(if2) )
> {
> currentDate <- time(if2)[i]
>
> equity<-getEndEq(b.strategy, currentDate)
> Posn <- getPosQty(b.strategy, Symbol='if2', Date=currentDate)
> #cat(as.character(i),"position on current bar is ",Posn, append = FALSE)
>
> if(!is.na(MA[i]))
> {
> if( Posn == 0 & signal[i] == 1 ) { #no marketposition, and long signal
> occurs
> #long entry
> openprice <- as.double((Op(if2[currentDate])))
> unitsize <- abs(as.numeric(trunc(equity/(openprice*300*0.15))))
> addTxn(b.strategy, Symbol='if2', TxnDate=currentDate,
> TxnPrice=openprice, TxnQty = unitsize ,
> TxnFees=-openprice*300*0.00005*unitsize, verbose = F)
> }
> else if( Posn != 0 & signal[i] == -1 ) {
> #exit position
> openprice <- as.double((Op(if2[currentDate])))
> unitsize <- abs(getPosQty(b.strategy, Symbol='if2',
> Date=currentDate))
> addTxn(b.strategy, Symbol='if2', TxnDate=currentDate,
> TxnPrice=openprice, TxnQty = -unitsize ,
> TxnFees=-openprice*300*0.00005*unitsize, verbose = F)
> }
> }
>
> updatePortf(b.strategy, Dates = currentDate)
> updateAcct(b.strategy, Dates = currentDate)
> updateEndEq(b.strategy, Dates = currentDate)
> }
>
> chart.Posn(b.strategy, Symbol = "if2", theme = myTheme, TA =
> "add_SMA(n=15,col=4)")
>
> txns <- getTxns(Portfolio = b.strategy, Symbol = "if2")
> tstats <- tradeStats(Portfolio = b.strategy, Symbol = "if2")
> [/code]
>
> data format is as below
>
> [code]
>
> 2012/1/18 9:15:00 2577.6 2580.6 2573.6 2577.2
> 2012/1/18 9:16:00 2577.2 2583.6 2577 2583.2
> 2012/1/18 9:17:00 2583.4 2583.8 2580.6 2580.8
> 2012/1/18 9:18:00 2580.8 2581.4 2576.4 2577
> 2012/1/18 9:19:00 2577 2582 2576.6 2580.4
> 2012/1/18 9:20:00 2580.4 2581.4 2578.2 2581
> 2012/1/18 9:21:00 2580.8 2580.8 2578.8 2580
> 2012/1/18 9:22:00 2580.2 2581 2579.4 2579.8
> 2012/1/18 9:23:00 2580 2584.4 2579.8 2583.6
> 2012/1/18 9:24:00 2583.8 2585.2 2576 2576.2
> 2012/1/18 9:25:00 2577 2579.8 2576.8 2579
> 2012/1/18 9:26:00 2578.8 2584.2 2578.8 2581.2
> 2012/1/18 9:27:00 2581 2582.2 2580.6 2581.6
> 2012/1/18 9:28:00 2581.6 2581.6 2578.8 2579.4
> 2012/1/18 9:29:00 2580 2583 2579.8 2581.4
>
> [/code]
>
>
>
> --
> View this message in context: http://r.789695.n4.nabble.com/Specify-time-segment-to-calculate-indicator-and-trade-tp4705639.html
> Sent from the Rmetrics mailing list archive at Nabble.com.
>
> _______________________________________________
> R-SIG-Finance at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-finance
> -- Subscriber-posting only. If you want to post, subscribe first.
> -- Also note that this is not the r-help list where general R questions should go.
>
More information about the R-SIG-Finance
mailing list