The findings from today’s Ipsos MRP poll show in no uncertain terms just how much trouble the Conservatives are in. Our model has the Conservatives winning just 115 seats, with Labour on 453. To compare, in 1997 Tony Blair’s Labour party won 418 and John Major’s Conservatives 178. On these results, the Conservatives could be heading for their worst general election defeat in modern political history. There is often a lot of mystery surrounding multi-level regression and post-stratification (MRP) in terms of what it is and how it works. In simple terms, we run a large-scale survey of almost 20,000 people on our online KnowledgePanel, recruited according to gold-standard random probability methods. This gives us a granular picture of how different demographic groups are likely to vote across the country. Then, because we know what proportion live in different constituencies, we can project the probabilities for how individual seats will vote, within certain confidence ranges. One of the key benefits over a normal poll is that the geographic detail helps us understand the pattern of current voting intentions. And we have produced a whole range of interactive maps for fellow geeks to get lost in. Looking at that detail explains why the Conservatives are performing so badly nationwide. Rishi Sunak’s party is losing support across the country, but especially in stronger areas such as the east and south of England and across the Midlands. Meanwhile, Labour is making more modest advances in its heartlands in highly urban areas, but bigger gains in Scotland and the north-east. Support for the Conservative party is falling most where it was strongest in 2019. Sometimes referred to as a “proportional swing”, this means it is losing more seats in the first past the post system than if the swing against it was more uniform. There is still some debate about what will be the final nature of the swing on election day – whether it will be more proportionate or more uniform – but the pattern shown in the raw survey data could be highly damaging for the Conservatives, as it was in the local elections and recent byelections such as Wellingborough. The Conservatives also find themselves challenged on both flanks by other parties. It has been manifesto week for Reform UK and our model shows them winning three seats, including Nigel Farage’s in Clacton. Our model also shows the Liberal Democrats taking seats off the Conservatives in the south of England. The top four issues according to the public have been clear for some time – the NHS, the economy, cost of living and immigration – and across all of them public dissatisfaction with the government’s record is deep. The Conservatives’ brand image as a party fit to govern has more than halved since 2019, and it can no longer rely on having a leader who is preferred to the opposition – Sunak is in the worst position of any prime minister starting an election campaign in Ipsos records going back to the 1970s. The public wants a change – even if they are not completely enamoured by Labour – and the Conservatives are finding that tide of opinion difficult to shift. We should always remember that polls and models such as this are snapshots of public opinion and not predictions of what is to come – we need to allow for some uncertainty in our interpretation. Things can change between now and polling day, but what is clear is that the Conservatives are losing support across the country, in different directions, and are in a deep political hole. Time will tell to what extent they can dig their way out, if they can at all.