The polls were accurate in 2016, showing Hillary winning the popular vote by 2 or 3 points which she did. There were a few bad state polls, such as Penn, OH (which showed a close race with Trump winning, but which Trump won by 9 or 10 points), WI, MI - but all those polls were close and Trump won MI, WI and Penn by a total of only 77,000 votes. Most state polls were accurate...AZ showed a close win for Trump, which he won by 4 or 5 points. It's easy to scream bad polls, but they weren't as bad as people want to believe.
Remember, margin of error works both ways. If a poll shows Trump ahead by 48 to 45 and the margin of error is three points...then that means Trump's support could be as low as 45 or as high as 51. Likewise, the dem could be as low as 42 or as high as 48.
While you may have a point, there is no denying that The averages are deliberately gamed by certain malevolent mainstream media outfits to suppress Republican votes. And while the final national polls may not have been too far off, the polls leading up to polling day are were skewed heavily for the Dems. They usually make the polls closer just before elections so they don’t look too bad.
For the midterms, you’d be better off taking your polls from (funny enough), The New York Times. Theism polls are transparent, era current and sometimes live.
https://www.nytimes.com/interactive/2018/upshot/elections-polls.html