Canada federal election 2008Liberal4.1 points behind → 26.3%
US President (national vote) 1960Richard M. Nixon (Republican)4.0 points behind → 49.5%
Spain general election 2016Sumar6.3 points behind → 21.1%
Canada federal election 2011Liberal6.3 points behind → 18.9%
Netherlands Tweede Kamer 2023PRO3.6 points behind → 15.8%
UK general election 2005Conservatives6.4 points behind → 33.2%
Italy Chamber 2013PdL6.6 points behind → 21.6%
Spain general election 2015PSOE3.2 points behind → 22.0%
France presidential, round 1 2007PS3.2 points behind → 25.9%
UK general election 1983Labour6.8 points behind → 28.3%
UK general election 1987Labour3.0 points behind → 31.5%
UK general election 1970Labour7.0 points behind → 43.9%
The biggest comebacks from 6 months out
Liberal 1993−26.0
Liberal 2025−19.1
Conservative 2006−9.7
George H. W. Bush (Republican) 1988−9.0
Conservatives 1974−7.8
Labor 2025−7.1
Conservative 1988−6.5
Conservative 2021−6.0
Bill Clinton (Democrat) 1992−6.0
Richard M. Nixon (Republican) 1968−5.0
How this works
For every election we hold polls for, we take each party's poll average at fixed points before polling day (two weeks, a month, two months, and so on to three years) and its gap to the party leading that average. Only the leader and the runner-up count: a comeback means the second party overtaking the first. A party "finished first" if it won the most votes, which is not always office: Gore in 2000 and Clinton in 2016 won the popular vote. "That far behind" means within two points either side of the figure you chose.
The data: Great Britain 1950–2024 (Mark Pack's PollBase; February 1974 only, as the two 1974 elections share a year); Australia, Canada, France (first round), Germany, Italy, the Netherlands, Poland and Spain from our own poll archive; US Senate, governor and presidential races state by state since 2000; the US national vote since 1952 (before 2000, Gallup's June poll only). Past comebacks are not a forecast: our forecast for each country already allows for how much polls move this far out.