[R-SIG-Finance] Blotter - Setting up a futures_series

Wolfgang Wu wobwu22 at yahoo.de
Tue May 11 15:29:58 CEST 2010


The best I could come up with is to write a quick function to check if it is possible to call updatePortf() function. To do this I test if each symbol in the portfolio has a row for the passed in date.

#function to test if you can get the value of the portfolio
fCanValue <- function(Portfolio, Date){
    canValue <- TRUE;
    tempInstr <- names(fPortfolio);
    for (i in 1:length(tempInstr)){ 
        xtsInstr <- get(tempInstr[i])
        if (length(index(xtsInstr[Date])) == 0) {canValue = FALSE}
    }
    return(canValue);
}

This might work if one of the symbols might be missing a day or so because of a holiday. I am then just not calling updatePortf().


How is the best way to solve the general problem that updatePortf() seems to assume that there is a row for the date you pass in for every symbol?Is the best route to solve this 
a) to remove the symbol from the portfolio when there are no more prices? Can I just delete the xts table in portfolio and the profit and loss calculation will still work?
b) ignore symbols that have 0 position and only produce an error when you have a position in a symbol but no price? There might be still problems with holidays and missing data.
c) try to use the last available price to mark a symbol?

Regards,
 
Wolfgang Wu



----- Ursprüngliche Mail ----
Von: Brian G. Peterson <brian at braverock.com>
An: Wolfgang Wu <wobwu22 at yahoo.de>
CC: R SIG Finance <R-SIG-Finance at stat.math.ethz.ch>
Gesendet: Donnerstag, den 6. Mai 2010, 11:46:14 Uhr
Betreff: Re: AW: [R-SIG-Finance] Blotter - Setting up a futures_series

Wolfgang Wu wrote:
> It seems like the error occurs when you run updatePortf() for a date where one of the symbols are not defined. So TE_N9 doesn't have a row for the date '2009-01-02' which is causing this error. I guess this will happen often as there might be holidays or non trading days which might be different for different symbols (maybe trading on different exchanges, etc..).  Are there any work-arounds to this problem?

That sounds like a bug.  I'll check and fix it to make sure updatePortf won't try to update a particular instrument beyond it's last traded date.  Ideally, the historical instrument data will be tested with 'try' and a warning rather than a fatal error will be generated.  I'll try to get to it tomorrow or over the weekend.

> Also is it possible to add or remove symbols to a portfolio object later than initPortf? 

On the list of functions to write...  contributed functions welcome.

Regards,

   - Brian

> Thanks
> 
> 
>  Wolfgang Wu
> 
> 
> 
> ----- Ursprüngliche Mail ----
> Von: Wolfgang Wu <wobwu22 at yahoo.de>
> An: Brian G. Peterson <brian at braverock.com>; R SIG Finance <R-SIG-Finance at stat.math.ethz.ch>
> Gesendet: Donnerstag, den 29. April 2010, 14:44:32 Uhr
> Betreff: Re: [R-SIG-Finance] Blotter - Setting up a futures_series
> 
> Sorry for being a pain in the neck but I've got another problem
> 
> When running the folloing code I get the error when I call updatePortf(): Error in if (!all(i <= 0)) stop("only zeros may be mixed with negative subscripts") :   missing value where TRUE/FALSE needed
> 
> #Example
> library(FinancialInstrument)
> library(blotter)
> currency("USD", multiplier=1);    future("TE", "USD", 1000)
> future_series("TE", "M9", "2009-01-02", "2009-01-03")
> future_series("TE", "N9", "2009-01-03", "2009-01-04")
> 
> TE_M9 <- xts(rbind(cbind(54.28, 55.92, 53.57, 55.73, 78959, 1), cbind(55.50, 55.70, 54.38, 55.63,  75464, 2)), order.by=as.Date(c("2009-01-02", "2009-01-03")))
> colnames(TE_M9) <- c("Open","High","Low","Close","Volume", "OpenInterest");
> 
> TE_N9 <- xts(rbind(cbind(54.28, 55.92, 53.57, 55.73, 78959, 1), cbind(55.50, 55.70, 54.38, 55.63,  75464, 2)), order.by=as.Date(c("2009-01-03", "2009-01-04")))
> colnames(TE_N9) <- c("Open","High","Low","Close","Volume", "OpenInterest");
> 
> initPortf(name='fPortfolio', symbols = c('TE_M9', 'TE_N9'), initPosQty = 0, initDate='2009-01-01')
> initAcct(name='fAccount',portfolios='fPortfolio', initDate='2009-01-01', initEq=10000000);
> addTxn('fPortfolio', Symbol='TE_M9', TxnDate="2009-01-02", TxnPrice=55, TxnQty=100, TxnFees=0, verbose=TRUE);
> updatePortf(Portfolio='fPortfolio', Dates = "2009-01-02");
> #-----------------------------------------------
> 
> Any idea what I am doing wrong or where the problem is?
> 
> Thanks!!
> 
> Wolfgang Wu
> 
> 
> 
> ----- Ursprüngliche Mail ----
> Von: Brian G. Peterson <brian at braverock.com>
> An: Wolfgang Wu <wobwu22 at yahoo.de>; R SIG Finance <R-SIG-Finance at stat.math.ethz.ch>
> Gesendet: Donnerstag, den 29. April 2010, 12:21:36 Uhr
> Betreff: Re: AW: [R-SIG-Finance] Blotter - Setting up a futures_series
> 
> On 04/29/2010 02:15 AM, Wolfgang Wu wrote:
>> Understood.
>> 
>> But in order to mark the book at some point you will need historical data of the contracts. How is the link from the contract specifications, using the future() and future_series() functions to the actual
>> historical data. What would be the equivalent to the example of a stock like this:
>> 
>> #Defined the stock as financial instrument
>> stock("GSPC",currency="USD",multiplier=1)
>> #Get the historical data for variable with name "GSPC"
>> getSymbols('^GSPC', src='yahoo', index.class=c("POSIXt","POSIXct"),from='1998-01-01')
>> 
>> Assuming I have the time series for CLM9 and CLN9 in an xts format
>>    
> 
> 
> FinancialInstrument uses an underscore to separate the root symbol (primary_id) from the series symbol (suffix_id).  In your example, this means that your historical contract series data should be named
> 
> CL_M9
> CL_N9
> 
> So that the blotter code can find them (blotter follows the quantmod convention of naming market data for the instrument)
> 
> you can see this by doing
> 
> ls(pos=.instrument)
> 
> which will show you the environment where the instrument definition objects are kept.
> 
> Regards,
> 
>    - Brian
> 
> 
>> CLM9
>>              Open  High   Low Close Volume OpenInterest
>> 2009-03-23 54.28 55.92 53.57 55.73  78959       187205
>> 2009-03-24 55.50 55.70 54.38 55.63  75464       190182
>> 2009-03-25 55.35 55.58 53.50 54.19 119624       198422
>> ...
>> 2009-05-19 59.16 60.48 58.55 59.65  40957         1838
>> 
>> and
>> 
>> CLN9
>>              Open  High   Low Close Volume OpenInterest
>> 2009-04-22 50.95 51.14 49.72 50.70  95249       178023
>> 2009-04-23 50.38 51.38 50.09 51.07  90781       181605
>> 2009-04-24 51.13 53.09 50.40 52.94  70061       187690
>> ...
>> 2009-06-22 69.89 69.89 66.25 66.93  32070         4299
>> 
>> Regards,
>> 
>> 
>>   Wolfgang Wu
>> 
>> 
>> 
>> ----- Ursprüngliche Mail ----
>> Von: Brian G. Peterson<brian at braverock.com>
>> An: Wolfgang Wu<wobwu22 at yahoo.de>
>> CC: R SIG Finance<R-SIG-Finance at stat.math.ethz.ch>
>> Gesendet: Donnerstag, den 29. April 2010, 0:28:48 Uhr
>> Betreff: Re: [R-SIG-Finance] Blotter - Setting up a futures_series
>> 
>> Wolfgang Wu wrote:
>>    
>>> Hello,
>>> 
>>> I am playing around with blotter and am trying to set up several futures contracts. So assume I've got data for the crude oil future CL. I want to set up futures for the June 09 and July 09 contract.
>>> What I've done is the following:
>>> 
>>> currency("USD", multiplier=1);    future("CL", "USD", 1000, identifiers=c("CLM9", "CLN9"))
>>> future_series("CL", "M9", "2009-03-23", "2009-05-19")
>>> future_series("CL", "N9", "2009-04-22", "2009-06-22")
>>> 
>>> 
>>> 1.) How do I link the future to the future_series? I assume this is by using the primary_id, correct?
>>>      
>> correct
>> 
>>    
>>> 2.) What do I specify in the identifiers in the futures function?
>>>      
>> anything or nothing at all.  identifiers are for other symbols, e.g. Bloomberg, Reuters, etc.  In your example, I would probably not add identifiers to the main contract.
>> 
>>    
>>> 3.) What is the multiplier in the future mean? Is it the value for one tick or for one point? So for oil one tick is 10 USD and one point move is 1000 USD.
>>>      
>> 'point value' in your terminology.  It is used to get from the price to the notional value in the listed currency.  I'll probably add tick value to the model later this week.
>> 
>>    
>>> Thanks!!!
>>>      
>> You're welcome.
>> 
>> Regards,
>> 
>>     - Brian
>> 
>>    
>>> Wolfgang Wu


-- Brian G. Peterson
http://braverock.com/brian/
Ph: 773-459-4973
IM: bgpbraverock





More information about the R-SIG-Finance mailing list