Another big problem with pre-election fraud was all the wrong polls. Those polls do not just suppress/increase voter turnout (my guy is going to lose -- no need to vote; my guy is winning -- I'll be glad to wait in line to vote for him just to add my voice).
Polls are also used by parties and candidates to raise funds. In SC, Sen Graham's challenger was predicted by polls to be even with Graham. The challenger and his party (D) used the polls to raise funds for the campaign. Nobody wants to waste their donations on a campaign that can't possibly win. The money, in turn, helps the candidate's changes. Self-licking ice cream cone.
The outcome? Even though Jaime Harrison raised a record $100M in the election cycle, Graham was called as the winner with just 46% of the vote reporting. With nearly all votes counted in the race now, Graham 54.5%, while Harrison received 44.2%. Not exactly the "toss-up" fund raisers and pollsters were predicting.

10.3 points is not a small number. That's a difference of 258,321 votes.
This is why polling will continue to favor Democrats. It interferes with election day voting, and dishonestly convinces donors to write checks.