[R] BradleyTerry "subscript out of bounds"

Jeff Newmiller jdnewmil at dcn.davis.ca.us
Fri Dec 29 11:00:34 CET 2006


I don't see the problem with the following... the citations and
baseball data work fine, but my simulated data seems to give
BTm a headache. What am I missing?

---
library(BradleyTerry)
library(doBy)
ng <- 100
players <- factor( sort( c( "jeff", "mike", "paul", "rich" ) ) )
np <- length( players )
p1 <- factor( c( rep( "jeff", ng )
                , rep( levels( players ), np ) )
             , levels=players )
p2 <- factor( c( rep( "mike", ng )
                , rep( levels( players ), each=np ) )
             , levels=players )
p1s <- c( rnorm( n=ng, mean=-0.5, sd=2 ), rep( 1, np^2 ) )
p2s <- c( rnorm( n=ng, mean=0.5, sd=2 ), rep( 0, np^2 ) )
results <- data.frame(
     winner=factor( levels(players)[ ifelse(p1s<p2s,p2,p1)], levels=players )
   , loser=factor( levels(players)[ ifelse(p1s<p2s,p1,p2) ], levels=players)
   , Freq=c( rep( 1, ng ), rep( 1, np^2) ) )
results <- summaryBy(Freq ~ winner + loser, data=results, FUN=c(sum) )
attr(results, "names") <- c( "winner", "loser", "Freq" )
results.btm <- BTm( results ~ .. ) # this spits out an error:
----
Error in BTm(results ~ ..) : subscript out of bounds

 > sessionInfo()
R version 2.0.1, 2004-11-15, i386-pc-linux-gnu

attached base packages:
[1] "methods"   "stats"     "graphics"  "grDevices" "utils"     "datasets"
[7] "base"

other attached packages:
BradleyTerry         brlr         doBy        Hmisc
      "0.8-5"      "0.8-8"        "1.6"      "3.0-1"

I have also tried this on Windows XP with R version 2.3.1 with same result.
-- 
---------------------------------------------------------------------------
Jeff Newmiller                        The     .....       .....  Go Live...
DCN:<jdnewmil at dcn.davis.ca.us>        Basics: ##.#.       ##.#.  Live Go...
                                       Live:   OO#.. Dead: OO#..  Playing
Research Engineer (Solar/Batteries            O.O#.       #.O#.  with
/Software/Embedded Controllers)               .OO#.       .OO#.  rocks...1k



More information about the R-help mailing list