[R] about yesterday predicted values

[Luis Silva]

I don't understand where is my problem. I tried the problem 
suggested by Prof. Brian Ripley

data(USAccDeaths, package="ts")

and it worked. So this is not a machine or configuration 
problem. This is my series (dataset1.ts)

          Jan      Feb      Mar      Apr      May      Jun  
Jul      Aug
1998 20957656 22280838 28048581 32286754 36456729 35575315 
34127179 34363108

1999 20597673 22495503 32313913 34403642 39324570 38507480 
36640790 36217094

2000 24841914 28977013 34243366 38180015 49119602 43809914 
41778152 45625301

2001 30366925 28705775 33911934 38647902 48854117 41389555 
43607174 41367514

2002 29111032 28055234 32778200 43894044 48513923 42018619 
38626430 38595853

2003 25680114 
27112446                                                      

          Sep      Oct      Nov      Dec
1998 34255786 33152873 29587196 25815553
1999 38141741 36711756 36950693 26657417
2000 46213806 45460357 39832263 31634397
2001 38920451 41416886 36272139 26559226
2002 42068682 41827283 33738721 31053206

I made 


dataset1.arima<-arima(x = dataset1.ts, order = c(2, 1, 0), 
seasonal = list(order = c(1, 1, 0)))

> predict(dataset1.arima,n.ahead=10)
$pred
               Mar           Apr           May           
Jun           Jul
2003  0.0007636733 -0.1296421510 -0.2341768144 -0.1240727672 -
0.1625043897
               Aug           Sep           Oct           
Nov           Dec
2003 -0.1726880234 -0.1537146941 -0.1628472385 -0.3301221034 -
0.2691800319

Where is my problem?
--
SAPO ADSL.PT, apanhe já o comboio da Banda Larga. Kit SAPO ADSL.PT €50

hTTP://www.sapo.pt/kitadsl