I see Trump as having an extremely good chance of winning the electoral college, possibly in a landslide, and an even chance of taking the popular vote.
First, as to electoral college prospects, I think you have to decide state by state whether Trump will run better in the election than in the polls and by how much. The polls in 2016 gave Hillary a 5-10 point lead in Wisconsin- enough for her to decide it was unnecessary to campaign there. The “shy Trump voter” effect was real. Even beyond that there are serious questions about systematic error in the polls. Who has a landline? Who answers their cellphone? Who has time to respond to a poll? When I give Trump the same extra margin he achieved in 2016, I am hard pressed to name a state Trump won in 2016 that he will lose in 2020. In Wisconsin, Trump is behind Biden by 2.6 points. It is not enough. He was behind Hillary by over 5.0
Second factor is that Trump will likely do better with Blacks, Asians, and Jews than last time. Not by much but by roughly 5 points. Nominally Democrat White urban and suburban women are wild cards. If they decide Democrats are to blame for rising homelessness, filth, and crime, they might just fail to pull the D lever. My overall point is that the electorate is not a static blob. Each segment is in play and is moved by different factors. People who disapprove of Trump may not necessarily cast a vote for one of his opponents.
A final factor is that we don’t know which side will cheat more effectively and will do better at preventing the other side from cheating. It is difficult to stop vote harvesting schemes. A lot can be done with absentee ballots.
Another way to think of it. Try to estimate the number of Trump 2016 voters who have changed their minds versus the number of non-Trump voters amenable to switching. My impressions is that more than a few Democrats have at least decided Trump is not so bad (e.g. #walkaway) or are disgusted by the dramatic jumps to the Left by the Democratic party. The election is far from decided.