I follow United States electoral politics fairly closely, which gives me somewhat of an advantage when discussing elections, but can also be a disadvantage because I have to remember that some things that I would consider obvious are obscure pieces of knowledge. I know what states went to which candidates in every election for the past 40 years, and have looked at electoral trends down to the county level. Understanding these trends does a lot to explain allegations of fraud in this election. So let me present a basic birds-eye view of the United States' electoral map, as it has existed since either 1988 or 1992.
Starting in 1988, but clearly by 1992, the United States' electoral map was a contrast between the Democratic Pacific Coast and Northeast Atlantic, with some of the midwest being a "swing area". From 1992 to 2012, Wisconsin, Michigan and Pennsylvania voted for the Democratic candidate, although sometimes by small margins. These states were about as Democratic as the national popular vote, or slightly more so. For example, even in 2004, the only election in the past seven where the Republican candidate won the popular vote, all three states voted Democratic. For example, in a nationwide vote that Bush won by 2.5 percentage points, John Kerry won Pennsylvania by 2.5 percentage points, making it 5% more Democratic-leaning than the nation as a whole. This long string of electoral successes in the Upper Midwest might have contributed to a sense of hubris in 2016, where there was a talk of these states being a "Blue Wall". In a national popular vote that Hillary Clinton's campaign won by 2 percentage points, she lost these three states by .7%, .3% and .3%, respectively. Forgive me for getting into the nitty gritty of these results, but the summary is: after two or three decades of being 2 to 5 points more Democratic than the country as a whole, these states became 2 to 3 points less Democratic than the country as a whole. This is a normal thing that occurs in elections, and is attributable to many things: demographic trends, candidate quality, local issues and campaign schedules, and other things that move the needle ever so slightly. Despite being slightly less Democratic leaning than the country as a whole in 2020, these states were still in line with their past results, and Biden's wins in them, while not absolutely predictable, are not at all surprising.
So it comes to a question of Occam's Razor. What is more likely: that these states that were traditionally Democratic leaning for six election cycles and then went narrowly to a Republican candidate for one election reverted to their mean, or that there was a vast conspiracy to rig the election? And if Wisconsin (for example), could only vote for a Democrat in 2020 because it was rigged, why did it vote for Democratic candidates in 1988, 1992, 1996, 2000, 2004, 2008 and 2012, as well? Especially since four of those elections were under the administrations of Republican Presidents. The justice departments of Ronald Reagan, George HW Bush and George W Bush all were bamboozled by the Democratic fraud machine in Wisconsin, but for some reason, under Barack Obama in 2016, the Democratic Party forgot to rig the election. But then, somehow, in 2020, under Trump, the Democratic Party reinvigorated the rigged election. To make the theory work, more and more epicycles need to be added.
It comes down to a simple question of what is more likely: states that had a historic voting pattern stayed within that historic voting pattern, or that there is a vast, decades-long conspiracy to rig votes in a few key states.
But let me get to the meta behind what is going on here. All that stuff I just wrote is something that someone could go and check fairly easily. It takes about 60 seconds to open up US Election Atlas or wikipedia and look at election results from 1988 or 2000 or any other year. Most of these people are not stupid. But they are pretending to be stupid. As long as they are acting dumb and pretending not to understand basic facts like "an incumbent with a high unemployment rate is always at a risk of losing", it then puts the burden on us, the people who are trying to be reasonable and fair, to refute arguments that are, from start to finish, bullshit. People aren't stupid, but by pretending to be stupid, we have to translate everything we say downwards into dumb terms to satisfy their tantrums. Perhaps I have fallen into the trap by even writing this.
So, in summary: Joe Biden won the election because he took advantage of voting trajectories in key states that have existed for decades, and that we are not required to answer every bad faith question from people who aren't bothering to access easily available data.