lections aren’t truly over until a loser concedes. And democracies aren’t truly democracies until executive authority is transferred by – or from – a losing incumbent.

With narrow wins in Wisconsin, Michigan and Pennsylvania, Trump won the Electoral College in 2016, defying opinion polls that had him behind in those states by four to seven points.
In the end, the 2016 election was decided by less than 78,000 votes, spread across those three states, just 0.08 percent of all votes cast. It is likely that 2020 will be decided by a similarly infinitesimal sliver of votes, perhaps coming from just one or two final states.
No one expected this election to be this close. Over the course of the campaign, I repeatedly urged caution in interpreting swing-state polls, unsure that the underestimation of Trumps strength in Midwestern states had been eradicated, that the election was much closer than polls were suggesting.
Elections aren’t over until someone concedes. 
Poll averages had Biden up by 8.4 points in Wisconsin; he currently leads by 0.6 points. In Michigan, the final poll average was Biden by 7.9 points versus an actual lead of 2.6 points. These are not small misses.
My cautionary advice to colleagues and journalists was that if 2016 poll error was replicated, Biden would probably win, if narrowly. That may come to pass.
But no matter the outcome, Trump has electoral strength that has been vastly underestimated again, along with Trumps hold on the Republican Party.
Incumbent Republican Senators in both of the Carolinas, in Maine, Iowa, and Kentucky had staggering sums of money thrown against them, literally hundreds of millions of dollars. All were returned easily.
Democratic hopes of flipping state legislatures in Pennsylvania, North Carolina, Michigan, Iowa and Texas all evaporated, along with hopes of control of those states redrawing of Congressional district boundaries in 2021.
Gerrymandering is here to stay. Democratic gains are therefore are at the state level where gerrymander has no influence picking up two Senate seats (Arizona and Colorado).
Trump has added the polls to his list of grievances about the likely result. Calling out specific polling firms and their errors, Trump described polls under-estimating his electoral strength as forms of election interference, as [voter] suppression polls, designed to demobilise his base of supporters.
But one should never infer conspiracy nor impute malice when sloth and incompetence suffice.
Trump has also promised a lot of litigation, alleging voter fraud and improprieties in vote counts. But Trumps Friday press conference had a tone of resignation to it, a listing of grievances more than a call to arms.
Election law experts are generally sceptical about the prospects for litigation to reverse Bidens provision and soon to be official wins in Wisconsin and Michigan, or the wins Biden may eke out in Pennsylvania or Georgia. Litigation on Trumps behalf has not successful in stopping the vote counts thus far.
Election officials have ploughed on with their work, almost surely aware that once processed and tallied, and once vote counts are complete, the prospect of successful legal challenges rapidly diminishes. Michigan has finished counting and is now moving to its certification of the state-wide result.
Various states give candidates the right to recounts if the final tallies are sufficiently close; like all things to do with American elections, these state-wide recount requirements and procedures vary tremendously.
But the Trump campaign has signalled it will avail itself of recount provisions in Wisconsin, the margin there sitting inside the recount trigger threshold of 1 percent.
Trump and his supporters have openly talked of the Supreme Court intervening. This seems unlikely at this stage. Until a clear, constitutional claim can be made, state law and state courts are the first port of call in litigating US election matters.
The Federal courts have not been especially receptive to claims that constitutional questions are in play at this stage, or that state officials or state courts have acted contrary to state law.
The Supreme Courts intervention in the 2000 election ending vote counts and the certification process in Florida, ensuring George W Bushs win over Al Gore came well over a month after election day, with a conservative majority of the court holding that Bushs rights to equal protection under the law were being violated by the prolonged and (in their view) consideration of ballots of dubious validity.
The long, drawn-out back and forth in Florida state courts was itself a trigger for the Supreme Courts intervention and rapid resolution of Bush v Gore in 2000.
Just two or three days after the election, most experts are struggling to see the basis for a Bush v Gore type intervention from the Supreme Court.
I concur with the balance of expert opinion that what we are seeing is garden-variety vote counting. What is distinctive is the huge number of votes cast, along with the fact that mail and early voting were overwhelmingly utilised by Democrats.
A similar phenomenon was observed in the 2018 midterms, so notable that election specialists gave it a name: an Election Night red mirage followed by a blue shift. Biden slowly running down Trumps leads as votes are counted was predicted by analysts on both sides of politics.
If Trump does lose, a further stress test awaits American democracy. We are on the verge of this election being over, at least in terms of the counting of votes.
But elections arent truly over until a loser concedes. And democracies arent truly democracies until executive authority is transferred by or from a losing incumbent.