Why is sed regex grouping different from proper regex?

Question

I\'m struggling to understand why is sed messing up a simple regular expression.

I\'ve got the following input inside a file called test: