What validity do you put in polls? Presidential, sporting events etc...

I like to look at the questions asked and who the people selected for the poll. Do I really care how many illegal aliens wouldn't consider voting for Mr. Trump? Likewise do I care how many Conservatives think Secretary Clinton lied?
If asked correctly and to the right group, they can be helpful but of course ... Lies, damn lies and statistics is always in the back of my head.

I see polls as just a bias based on the sample size it was pulled from.

For example: Should Tom Brady be suspended from the Patriots for Deflategate?
New England Residents: Hell no... We love you Tom
Everyone else: Hell Yes... Cheating Tom and the Cheatriots... go back to cheaterville.

Most polls have some sort of bias to them already which is why they are selective on where they draw their responses from.

If you were doing an anti-gun poll, you'll probably draw your responses from liberals at Sandy Hook or Orlando. If you were pro-gun, you'd probably conduct your poll at an NRA headquarters or any place in Texas.

Polls generally have about as much validity as those Facebook quizzes which tell you which Hogwarts House you'd be sorted into.

The only poll that counts is the final vote...and lately, that is even now questionable. All the rest are media agenda drive drivel. They poll area's they know are friendly to their agenda or they just make crap up. They aren't not required to actually prove they did a poll.