Sed using extended regexp and capture groups

Question

I wrote the following sed one liner for substituting the 3rd portion of an IP address.

sed 's/192\.168\.[0-9]*\.\([0-9]*\)/192.168.15.\1/g'

192.168.0.1
192.168.15.1

I want to make it lazy so it would replace 192.168.0 with 192.168.15 by using ?? which is extended regexp but when I supply the '-r' option I get the following error.

$ sed -r 's/192\.168\.[0-9]*\.\([0-9]*\)/192.168.15.\1/g'
sed: -e expression #1, char 44: invalid reference \1 on `s' command's RHS

Can you explain what is going on here and how to make an expression lazy whilst using capture groups?

Note: I can achieve the example using many different approaches, my question isn't how to solve the example. I am interesting in using sed with exteneded regexp and capture groups.

Answer 1

There is no need to escape braces because ([0-9]*) is valid statement matching group of symbols.

$> echo "192.168.0.1" | sed -r 's/192\.168\.[0-9]*\.([0-9]*)/192\.168\.15\.\1/g'
192.168.15.1