Python is not built with math and statistics in mind, and this doesn’t work without using a package.
If you're looking to move from R to Python, here are two interesting posts: Python is Weird (an unabashedly biased intro to Python for R users); Programming with Data: Python and Pandas. The first one, from which the above quotation is extracted, provides a side-by-side comparison of some of the features of each language. You might like or not, since R is a DSL and Python is not a good PL to compare. The second one is a complete tutorial on Pandas (including linear regression) in IPython notebooks. Besides, Chris Albon's Technical Notes On Using Data Science & Artificial Intelligence To Fight For Something That Matters are also worth a look.