In a normal U.S. presidential election, the choice of vice-presidential candidate usually has only a minimal impact on the race. Vice-presidential candidates can help deliver their home state and provide a short-term boost to his or her running mate's campaign, but elections are rarely decided by who the "Veep" will be.
But this is not a normal election.
Donald Trump, Republican nominee for the U.S. presidency, announced Friday that Gov. Mike Pence of Indiana will be his pick for vice-president.
Hillary Clinton, the Democratic nominee, will make her pick before the party's convention in Philadelphia later this month.
Vice-presidential candidates are often chosen to fill a gap in the presidential candidate's resumé, be it geographic, demographic, or political. With Pence as Trump's running mate, it would give his ticket much-needed political experience, as well as a figure outside of the U.S. Northeast.
But vice-presidential picks normally do not have a significant impact on the race. The choice may help a presidential campaign at the margins, but the first rule of a vice-presidential candidate is often cited as "do no harm." Vice-presidential picks are at greater risk of dragging down a campaign than boosting it to new heights.
Clinton and Trump, however, are two candidates with the lowest favourability ratings in recent American electoral history. This may be a rare election in which the vice-presidential candidates can only help.
What kind of impact can we expect the picks to have on the state of the race?
'Veep' honeymoons fade fast
The examples of the last three U.S. elections suggest the impact a vice-presidential pick has on the polls can be significant — but it doesn't last.
In 2004 and 2008, the vice-presidential candidacies of John Edwards and Joe Biden coincided with a boost in the Democratic campaigns of both John Kerry and Barack Obama, who were only leading their Republican rivals by about a point when their picks were announced.
Within a week, Kerry was up on George W. Bush by three points and Obama was ahead of John McCain by more than six.
On the Republican side, VP choices in 2008 and 2012 initially helped their losing campaigns. With the choice of Sarah Palin, McCain went from a four-point deficit to a lead of three points a little over a week later.
In 2012, Paul Ryan helped push Mitt Romney's trailing campaign into a tie with Obama.
But after reaching a polling peak a week or two after the announcement of a vice-presidential running mate, in every case, these surging campaigns saw their numbers fall back down to Earth about as quickly as they had risen.
A large factor in this was the timing of the party conventions. Depending on when they were held, they tended to amplify the impact of a VP pick or put an end to an opponent's rising numbers — and sometimes both.
Based on these examples, we might expect the addition of Pence to the Trump ticket, as well as the Republican convention next week, to boost Trump's polling numbers by three or four points. But it might endure no longer than the beginning of the Democratic convention on July 25.
Veeps can be a drag
While many vice-presidential candidacies have had only limited effects on the longer-term chances of their running mates, the example of Palin shows how a bad VP pick can be disastrous.
When Palin, then the largely unknown governor of Alaska, was announced as McCain's pick on Aug. 29, 2008, the reaction was largely positive. A Gallup survey taken a week later found that 53 per cent of Americans had a favourable view of Palin. Just 28 per cent had an unfavourable view of her.
But by the end of the campaign, her favourables had dropped to 42 per cent and her unfavourables had soared to 49 per cent. An ABC News survey taken just before election day found that just 17 per cent of respondents said Palin's candidacy made them more likely to vote for McCain, compared to 46 per cent who said it made them less likely to vote Republican.
Biden, by comparison, was a net positive for Obama: twice as many people said it made them more likely to vote for him than less likely. Most Americans, though, thought Biden made no difference.
Palin, unfortunately for McCain, did.
If there is one thing that vice-presidents can do, it is help give their party a better chance of winning the Veep's home state.
In presidential elections over the last 40 years, non-incumbent vice-presidential candidates have seen a bigger swing towards their respective party in their home state than nationally.
In 2008, the Democrats increased their vote share by 8.5 points in Biden's home state of Delaware, compared to a gain of 4.6 points nationally. In 2012, the Republicans gained 3.6 points in Ryan's home state of Wisconsin, compared to a gain of 1.5 points nationally.
Ryan, however, failed to carry Wisconsin for the Republicans.
But on average since 1976, a vice-presidential candidate has boosted his or her party by 2.3 points in their home state, compared to the national swing.
That might make Pence an unusual choice for Trump. The Republicans are already leading in Indiana by a comfortable margin. Any boost Pence gives to the party in his home state would likely be wasted.
But as part of Trump's strategy to carry states in the Midwest, the choice would make more sense. Michigan and Ohio, two potential swing states, are just across the Indiana border.
Those states would help the Republican cause. But in this election, Trump may need more than a "do no harm" Veep.