Hillary Clinton won the popular vote, using 2016 to argue polls are massively off is basically admitting you don't understand what you're talking about.
The article lists other sources and a quick search will get you a lot further than wanting me to be AI to fulfill your every request. Point being they were wrong and surveys are not as reliable as fact and can be manipulated.
You're talking about completely different things. Odds are an estimate of who will come out over the line quicker, polls are the underlying data used to calibrate that model.
The odds set in 2016 by various groups were often wildly bearish on Trump winning, the polls were not so much. Political polls and surveys are usually not off the mark by all that much, especially if you have good demography data.
Surveys can be manipulated but they are also much better than flying blind with no data at all.
