The idea that Harry Truman was basically Donald Trump without Twitter is pretty funny. And the suggestion that the experts have it all wrong now because they did so in 1948 and 2016 kind of ignores the 16 presidential elections between those two meteoric events; in most, the prognosticators got it right.

But just for fun, let’s look at the analogy more closely and see if it holds out any particular hope for Trump fans looking for a big upset. Here are some factors that fed into the 1948 shocker that might distinguish them from today’s circumstances:

There Weren’t Many “Experts” in 1948, and They Weren’t Watching Closely

The political commentariat in 1948 was wildly different, and much smaller, than today’s, centered mostly in daily newspapers. And there were only three national polling operations (Gallup, Roper, and Crossley), all utilizing a primitive version of weighted sampling called “quota sampling” that had led to serious errors in prior elections. Gallup famously quit polling weeks out from the election after “finding” a five-point Dewey lead.

There was nothing in 1948 like today’s plethora of national and state pollsters, all constantly refining methodologies and being held accountable for accuracy and transparency. And no, they are not a bunch of chumps: The myth that pollsters all got 2016 “wrong” is largely itself wrong; national pollsters on average came reasonably close to the popular-vote results, won by Clinton by just over 2 percent. By contrast, 1948’s sparse polls missed a big swing to Truman, who won the national popular vote by 4.5 percent.

Dewey Was Immensely Overconfident

Every account of the 1948 contest describes Republican nominee Thomas Dewey as an overconfident candidate who was out-campaigned by Truman and took few if any risks. It was understandable. Democrats had won four straight presidential elections under FDR and were now running an incumbent who had only been elevated to the office by his predecessor’s death. The post–World War II atmosphere seemed much like post–World War I’s “return to normalcy,” which produced a Republican landslide. It was clearly “time for a change,” and the 1946 midterms showed so, with Democrats losing 54 House seats and ten Senate seats, along with control of both chambers.

Thanks to what happened in 2016, there is probably no Democrat in the entire country who is “overconfident” right now. Until a wooden stake is pounded into the 45th presidency, the possibility of another Trump threading-of-the-Electoral-College win, or even a post-election Trump refusal to accept defeat, will keep Democrats on their toes and alert to any negative trends. Whatever his weaknesses, moreover, Joe Biden has nothing of Tom Dewey’s chilly, stiff persona, which lent itself to perceptions he just didn’t want victory enough.