Fair points.  I agree with a lot of it.  For me "defund the police" is just a bad tagline.  "Reorganisation of law enforcement" is a lot less polarising, granted it is not as catchy.  People also vote on a single policy, that is true as well. 

I used to think people didn't know what they were voting for, but two things changed my mind.  The first being I have heard that every 2 years for over 20 years straight...  at some point one has to start wondering/questioning.  The other change was when I moved to the south.  I don't know where you live nor do I want to make any assumptions, but moving from the north to the south was eye opening.  I can't drive more than 100 ft without seeing a "Let's Go Brandon" sign (either yard or sticker on a car).  When I lived in the north I thought the media was accurate in cascading what the average person wanted/believed.  Now that I live in the south, yeah I don't anymore. 

Either way, I can compromise to a middle ground.  Fair, many don't know what they are voting for.  But I do think the "conservative minority" is larger than presented in media.  

I live in a fairly white conservative part of the midwest. Which admittingly has some very blue cities, but I've never lived in those cities.

There's a lot of gerrymandering here. In 2018, Democrats won the vote by over 8%, and yet only got 36 seats, whereas Republicans got 63, despite the massive win.

Sometimes it's more important to ask what people think they're voting against, rather than what they're voting for. (There's also a partisan factor that a lot of people love to vote for Trump, even if there was  Democratic politician with the same views, they wouldn't like as much.)

You are being a little disingenuous. And you should know better as a statistician. You can look at methodologies and what separates a bad poll from a good poll. You can critique individual polls too. The way and framing of questions is part of that. Just dismissing them outright though is weird.

And wait who said Hillary was supposed to decimate Trump? Everyone who looked closely at the elections was saying she's a bad candidate and it shouldn't have been her. Hell there was an entire scandal about the Clinton foundation using funding they gave to the DNC to make sure she won over Bernie. What people were saying was she would get more votes and she did. 

As for the question of people knowing or not knowing what they vote for, read manufactured consent. Also maybe look into the history of brainwashing and political maneuvering.  Messaging and perception have shown to be more effective than actual good policies people themselves want in countries where media literacy training is not taught. That's why teaching children how to get knowledge is way more important than maths, science, and all the things people thinks schools should be for. It's not that people are dumb, it's that we are people and people fall for things. Even the most intelligent of us, unless we are taught about and know how to avoid these tactics. Even then we fall for some of it. You and me and everyone here and everyone in the world.

Those with media literacy and taught critical thinking skills just fare better. Unfortunately those people are not the majority in the world because a properly educated population is a threat to the political careers of I would say most of the world politicians.

I'll take your two major points in one link. 

https://www.pewresearch.org/fact-tank/2016/11/09/why-2016-election-polls-missed-their-mark/

1) Hillary was predicted as the HEAVY favorite according to polls....   

When the United States of America got established, due to it's sheer size and not-yet-existing faster ways of transport and communication, the electoral system made much sense back then. But with the advent of trains and especially the telegraph, it had already become archaic even before the civil war started. And yet it is still in use today, with all it's flaws that make it's voting system very easy to play around.

There are several ways to solve the issue. I'll list some of the potential solutions here:

1. The easiest way: Make it a 2-step process. On the first round you can vote for whoever you feel like, and if neither gets past 50%, the 2 candidates with the highest results get into a runoff. It's a system that is in place in many other countries, and that way the spoiler effect would be eliminated, since there is nobody else on the second round, just the 2 candidates with the highest results in the first round.
2. Switch to a single transferable vote system. That way one can still vote for their preferred party without wasting their vote, as they can put the party (or parties) that they don't like to the bottom of the list. The biggest problem is that even in countries who have such a voting system, like Australia, not everybody knows how to vote in those ballots correctly, which can leave the door wide open to some shenanigans from parties who want to force some result.
3. Outlaw the winner-takes-all system. Winner-takes-all basically forces election over time to become 2-horse races due to tactical voting. Without it, minor parties would also get some chance at representation, which could defuse the increasingly partisan politics of the US, as the result could then very well be that neither gets past 270 electoral votes and thus would have to do a coalition with another party - and royally pissing off a potential coalition partner is simply not an option in such a type of government. It is also one where the electoral college could be a good thing to some degree, as a small party who doesn't have nationwide representation but instead focuses on one or a couple key states would have an easier chance to get representation than getting drowned out in a nationwide proportional election. For those who don't understand what I mean by that, look at the British elections and imagine that the SNP, Sinn Féin and Plaid Cymru would need to win the popular vote in all of the UK instead of just in their regions they're actually representing.
4. While we're at the topic of the electoral college, it needs to be either abolished, or, with a system like I explained at point 3 above, seriously overhauled. So let's get into ways on how the electoral college could be reformed: The first step: Remove the electors in the electoral college and simply count them as points that a party gets, which removes any possibility of unfaithful electors and the costs of sending hundreds of persons to D.C. just to cast a pointless vote. 2. The number of electoral votes hinges on the amount of senators and house representatives a state has, so a minimum of 3 (2 senators and 1 representative). So either this number gets reduced to a minimum of 1 with the same number of electoral votes, redistributing the freed up electoral votes proportionally by population, or increase the number of total electoral votes so that 3 would actually be representative to the population of the states with the lowest population (For comparison, elections for the European parliament have a minimum of 6, attributed to Luxembourg, Malta and Cyprus (though the latter should have more by now, as it's population rose rapidly in the last decade) and a maximum of 96 for Germany, which also has roughly 16 times the population of Malta or Luxembourg).

I could come with some more suggestions, but I'll leave it at that for now, as it's already quite the wall of text.

I don't get why pollsters are still throwing this myth around. Would they just have a good look at their own (and other pollsters numbers, and they'd know what's up:

While Trump won the electoral college, Hillary still won the popular vote by about 2%, or almost 3 Million people. Polls taken during the last 2 weeks before the election showed Hillary's support crumbling, fluctuating wildly between -5% and +7% of the popular vote, though most of them between 3% and 4%. While that's a bit higher than what she got in the end, it's still within the margin of error. Heck, if you take the mean value between the extremes, you come with just a measly 1% lead, even less than she carried in the end.

But in their calculations, they took the numbers of the entire last month, which includes also leads of up to 15%. Of course this skews the prediction very much in favor of Hillary despite those numbers being outdated. This is what went wrong with the polling, they looked too far back and had their predictions wrong due to that.

And you aren't concerned that in a span of 2 months her lead went from 15% to 4%?  Seems like a crazy fluctuation within 60 days.  Personally it makes me question how accurate the original 15% was. 

Look, 2016 simply made me question polls and media.  This whole conversation really started when Bandorr falsely and wrongly accused me of being a conservative conspiracy nut.  My point is really simple, because of 2016 I'm just not convinced our predictions are as accurate as we think they are.  It changed my perspective.

Lots people think this November there will be a bunch of SCOTUS backlash....  I'm not convinced.  Time will tell.    

Recent polls say otherwise.

People care more about their daily lives, and think Democrats aren't doing enough, or worse think Democrats are actively making things worse.

Fair enough, it is has been a bit since I looked at polls.

I think democrats will not do well this November.  Stock market is down, loan forgiveness isn't going over well, etc.  While I don't think democrats are necessarily at fault, I do think they will be blamed this month.  

538's forecast right now:

85% chance Republicans take the House.

53% chance Republicans take the Senate.