Updating A Large Table: Set @@rowcount Versus Explicit Transactions

When I need to perform an update against multi-million row table I
typically specify @@rowcount, to reduce locks.

e.g.
set @@rowcount 1000
while exists (select * from myTable where col2 is null)
update myTable
set col2 = col1 + 'blahblah'
where col2 is null



However, my boss' script does something like this. I think it works OK
but it seems overly complicated to me. Any thoughts?

while exists (....)
begin tran

insert into #table
select ...

update myTable
set ...
from myTable join #table ...

(@numberOfRows is a counter variable, tracking #rows that have been
updated since last batch)

if @numberOfRows > 1000
begin
commit
begin tran
end

end