Dictionaries in Python 3.6 are ordered, but that feature is considered an implementation detail that you shouldn't rely upon (except in a few specific cases like **kwargs). If you do require a specific order, you should use collections.OrderedDict instead. You can construct one using a list of key, value tuples that are in the desired order:
from collections import OrderedDict
finaldict = OrderedDict([('Visitor Team', visitor_team),
('Visitor Rating', visitor_rating),
('Home Team', home_team),
('Home Rating', home_rating),
('Expected Winner', expected_winner),
('Margin', expected_winner_diff),
])
An OrderedDict works just like a normal dict in most respects, other than having a different repr and a few additional methods. You can read more about it in the docs.
In Python 3.6+ you'd also be able to use keyword arguments to the constructor if your key strings were valid identifiers (e.g. OrderedDict(Margin=expected_winner_diff)). Unlike the ordering of normal dicts, the order of keywords is guaranteed to be preserved (not an implementation detail). That's not backwards compatible though (and can't work for your non-identifier keys anyway).
But it's probably worth considering that if you need a very specific order for your data, a dictionary may not be the best type to use to store it in. I see the tabulate function you're using comes from a library, and according to the docs, it accepts many different formats of data. I'd probably just pass it a list of column data, and give it the headers separately:
data = [visitor_team, visitor_rating, home_team,
home_rating, expected_winner, expected_winner_diff]
headers = ["Visitor Team", "Visitor Rating", "Home Team",
"Home Rating", "Expected Winner", "Margin"]
print(tabulate(data, headers=headers, floatfmt=".2f", tablefmt="fancy_grid"))
(Note, I've not actually tested that code, since I don't have the tabulate library on my system. But it should at least be close to working.)