Tossup in a competitive election.
Reasonable assessment -- with the caveat that the election be competitive. In a 50-50 election, Trump would have won New Hampshire.
Dubya barely won New Hampshire, and it could have been the difference between Gore and Dubya that year. Kerry somehow won it in 2004 (being from the neighboring state?). New Hampshire went by high single-digits in 2008 but was close to being the tipping-point state. In 2012 it was close to Obama's margin of victory. Hillary Clinton barely won the state despite winning the vote that did not matter.
It is easy to assume that because the last three Presidents have been re-elected with electoral maps similar to those with which they won that we can reasonably expect much the same in 2020. All three Presidents convinced 90-95% of those who voted for them that their vote was the right one and peeled little support from those who voted otherwise in their first Presidential election. Nationwide and state polling suggest otherwise this time. Approval for the President has been consistently below what is usual for a President having a good chance of being re-elected at similar stages of his Presidency (about 45% approval, which is about what Obama had... and as the record shows, needed it) until the government shutdown, which has cut further into the President's approval rating.
As someone who cares about this country and humanity I do not want to see President Trump's approval sink any lower. Such means that the economy has gone into recession -- or perhaps hundreds of thousands of people are dying in Seoul from a missile attack from North Korea.
His approval ratings will likely recover a little from the abysmal levels that he has reached in January 2019 as the federal shutdown comes to an end. But not much. A poll from this week shows that fewer than 40% of New Hampshire voters intend to vote for President Trump. Such is consistent with the President losing like Hoover in 1932 or Carter in 1980 -- if for very different reasons.