R help - Jul 2004 - Block Bootstrap Simulation

I am trying to run a time series block bootstrap.  I have a dataset of length t
= n*k.
I want to sample n blocks of length k from the data.  

I don't want to use loop because a) the overhead and b) I'm comparing it
with another technique that doesn't require a loop, and programming the
simulation without the loop gave me an insight into the solution of the problem
Is there a way of doing this without using a loop?

s<-1:n
v<-sample(s, replace=TRUE)
w<-v:(v+k)

Doesn't work

Phineas Campbell
www.phineas.pwp.blueyonder.co.uk

On 7/16/04 11:51 AM, "Campbell" <p.campbell at econ.bbk.ac.uk>
wrote:
> I don't want to use loop because a) the overhead and b) I'm
comparing it with
> another technique that doesn't require a loop, and programming the
simulation
> without the loop gave me an insight into the solution of the problem
> Is there a way of doing this without using a loop?
> 
> s<-1:n
> v<-sample(s, replace=TRUE)
> w<-v:(v+k)
> 
> Doesn't work
Why doesn't it work, ie if you replace w<-v:(v+k) by
w<-(v*k):((v+1)*k)?

ingmar