See this post for explanations of the metrics discussed below.
Rasmussen Reports publishes a presidential-approval poll that I have been recording since the fall of 2008, when Barack Obama was elected to his first term. One of the statistics that I glean from the polling results is what I call an enthusiasm ratio. Obama won re-election on November 8, 2012, in the 198th week of his presidency. His enthusiasm ratio in the week before the election hovered around 0.75. Trump’s enthusiasm ratio hasn’t been that low since the first year of his presidency:
On the other hand, there is the “right track”/”wrong direction metric”:
In the week before the election of 2012, the ratio stood at 0.8. That was as high as it had been since the beginning of Obama’s presidency, and a good omen for Obama’s re-election. The metric has hovered around 0.8 in recent months, but it should probably rise above 0.8 in the week before the 2020 election if one is to feel confident (or dismayed) about the prospect of Trump’s re-election.
Note: The slight discrepancy in the horizontal scales of the two figures is caused by the frequency of the underlying statistics: daily in the case of the first figure; weekly in the case of the second one. Converting days to weeks (as is the case with the first figure) causes the slight discrepancy. Specifically, 366 days/7 days per week = 52.29 weeks and 365 days/7weeks = 52.14 weeks, not 52 weeks. Over a span of 4 years, there’s a difference of 0.71 week between the two methods.