{"id":5722,"date":"2024-07-03T20:38:05","date_gmt":"2024-07-03T19:38:05","guid":{"rendered":"https:\/\/marriott-stats.com\/nigels-blog\/?p=5722"},"modified":"2024-07-03T21:06:29","modified_gmt":"2024-07-03T20:06:29","slug":"uk-general-election-2024-forecast-3","status":"publish","type":"post","link":"https:\/\/marriott-stats.com\/nigels-blog\/uk-general-election-2024-forecast-3\/","title":{"rendered":"UK General Election 2024 #5 &#8211; My Final Forecast as of 3rd July 2024"},"content":{"rendered":"<p>My third and final forecast of the number of seats to be won by each party in the <strong>2024<\/strong> UK General Election is &#8211;<\/p>\n<ul>\n<li><strong><span style=\"color: #0000ff;\">CON 130 <\/span><\/strong><span style=\"color: #0000ff;\">(+17)<\/span><\/li>\n<li><strong><span style=\"color: #ff0000;\">LAB 420 <\/span><\/strong><span style=\"color: #ff0000;\">(-17)<\/span><\/li>\n<li><span style=\"color: #ff6600;\"><strong>LD 50<\/strong><\/span><\/li>\n<li><strong>SNP 25<\/strong><\/li>\n<li><span style=\"color: #008000;\"><strong>GRN 3<\/strong><\/span><\/li>\n<li><span style=\"color: #339966;\"><strong>PC 2<\/strong><\/span><\/li>\n<li><span style=\"color: #00ccff;\"><strong>REF 2<\/strong><\/span><\/li>\n<li><strong>OTH NI 18<\/strong><\/li>\n<\/ul>\n<p>The numbers in brackets are the changes from <a href=\"https:\/\/marriott-stats.com\/nigels-blog\/uk-general-election-2024-forecast-2\/\" target=\"_blank\" rel=\"noopener\">my second forecast of 30th June 2024<\/a>.<\/p>\n<p><span style=\"color: #993300;\"><em>This article is based on polling data as of 2000 on 3rd July 2024.\u00a0 If more polling data comes in and causes me to change my forescript, the changes will be detailed in a postscript at the end of the article.<\/em><\/span><\/p>\n<p><!--more--><\/p>\n<h5><span style=\"color: #008000;\"><strong>My Articles on the 2024 UK General Election<\/strong><\/span><\/h5>\n<ol>\n<li><a href=\"https:\/\/marriott-stats.com\/nigels-blog\/uk-general-election-2019-2-my-forecast-beats-the-exit-poll\/\" target=\"_blank\" rel=\"noopener\">The most accurate forecaster of the 2019 general election<\/a> &#8211; I was independently assessed as the most accurate forecaster beating even Sir John Curtice&#8217;s exit poll.<\/li>\n<li><a href=\"https:\/\/marriott-stats.com\/nigels-blog\/uk-general-elections-keir-starmers-road-to-downing-street\/\" target=\"_blank\" rel=\"noopener\">Keir Starmer&#8217;s Train to Downing Street<\/a> &#8211; My assessment in 2021 of what Labour needed to do to win the next election.<\/li>\n<li><a href=\"https:\/\/marriott-stats.com\/nigels-blog\/my-election-forecasting-track-record\/\" target=\"_blank\" rel=\"noopener\">My election forecasting Track Record 2010 to 2024<\/a> &#8211; A list of all election forecasts I have made for General, European and Local elections, how I made them, how they turned out and what lessons I learned.<\/li>\n<li><a href=\"https:\/\/marriott-stats.com\/nigels-blog\/uk-general-elections-how-accurate-are-opinion-polls-revised\/\" target=\"_blank\" rel=\"noopener\">How Accurate are Voting Intention Polls (Revised)<\/a> &#8211; A recent article which explains why I now think the polls are accurate when before they would often underestimate the Conservative lead over Labour.<\/li>\n<li><a href=\"https:\/\/marriott-stats.com\/nigels-blog\/uk-general-election-2024-swing-and-turnout-forecast\/\" target=\"_blank\" rel=\"noopener\">Going Beyond the Swing in 2024<\/a> &#8211; A preliminary look at the 2024 general election at the start of the year.\u00a0 I give probabilities for 10 specific outcomes.<\/li>\n<li><a href=\"https:\/\/marriott-stats.com\/nigels-blog\/uk-general-election-2024-forecasting-model-gb\/\" target=\"_blank\" rel=\"noopener\">My 2024 UK General Election Forecasting Model<\/a> &#8211; I use a top down forecasting approach based on the sum and difference in Conservative and Labour vote share to predict share of seats won.<\/li>\n<li><a href=\"https:\/\/marriott-stats.com\/nigels-blog\/uk-general-election-2024-forecast-1\/\" target=\"_blank\" rel=\"noopener\">My 2024 UK General Election Forecast as of 23rd June 2024<\/a> &#8211; This explains how I use simulation to arrive at my estimate of <span style=\"color: #0000ff;\"><strong>CON 108<\/strong><\/span>, <strong><span style=\"color: #ff0000;\">LAB 442<\/span>,<\/strong> <strong>OTH 100<\/strong>.<\/li>\n<li><a href=\"https:\/\/marriott-stats.com\/nigels-blog\/uk-general-election-2024-forecast-2\/\" target=\"_blank\" rel=\"noopener\">My 2024 UK General Election Forecast as of 30th June 2024<\/a> &#8211; This is an update which explains how I arrived at my estimate of <span style=\"color: #0000ff;\"><strong>CON 113<\/strong> <\/span><strong><span style=\"color: #ff0000;\">LAB 437<\/span><\/strong> <strong>OTH 100<\/strong><\/li>\n<\/ol>\n<p>&nbsp;<\/p>\n<h5><span style=\"color: #008000;\"><strong>Data used in this article<\/strong><\/span><\/h5>\n<p>All polling data prior to April 2024 comes from <a href=\"https:\/\/www.markpack.org.uk\/opinion-polls\/\" target=\"_blank\" rel=\"noopener\">Mark Pack&#8217;s invaluable Pollbase<\/a>.\u00a0 For data since April, I am using <a href=\"https:\/\/www.bbc.co.uk\/news\/uk-politics-68079726\" target=\"_blank\" rel=\"noopener\">the BBC poll tracker<\/a> supplemented by <a href=\"https:\/\/en.wikipedia.org\/wiki\/Opinion_polling_for_the_2024_United_Kingdom_general_election#National_poll_results\" target=\"_blank\" rel=\"noopener\">the Wikipedia poll tracker<\/a>.\u00a0 For a summary of what the polls are saying as of 1600 3rd July 2024, <a href=\"https:\/\/x.com\/MarriottNigel\/status\/1808543302263976259\" target=\"_blank\" rel=\"noopener\">see this X\/Twitter thread<\/a>.<\/p>\n<p>All electoral data I display and use in this article comes from the <a href=\"https:\/\/commonslibrary.parliament.uk\/data-tools-and-resources\/parliament-elections-data\/\" target=\"_blank\" rel=\"noopener\">House of Commons Research Library<\/a>.\u00a0 The PDF file I use most of the time is this one <a href=\"https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/06\/UK-Election-Trends-1918-2019.pdf\" target=\"_blank\" rel=\"noopener\">UK Election Trends 1918-2019<\/a>.\u00a0 I&#8217;ve created a spreadsheet <a href=\"https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/06\/GB-General-Election-Data-1918-2019-votes-and-seats.xlsx\">GB General Election Data 1918-2019 &#8211; votes and seats<\/a> which contains all the data used to build my models.\u00a0 For clarity &#8211;<\/p>\n<ul>\n<li><strong>Seat Share<\/strong> is the percentage of seats in Great Britain won by each party.<\/li>\n<li><strong>National Vote Share<\/strong> is the percentage of all votes cast in Great Britain for each party.<\/li>\n<li><strong>Average Vote Share<\/strong> (per Seat) is the average of the vote share in each British constituency the party stands in.<\/li>\n<\/ul>\n<p>I first explained the national and average vote share concepts in my article &#8220;<a href=\"https:\/\/marriott-stats.com\/nigels-blog\/uk-general-elections-how-accurate-are-opinion-polls-revised\/\" target=\"_blank\" rel=\"noopener\"><em>How accurate are voting intention polls? &#8211; Revised<\/em><\/a>&#8220;.<\/p>\n<p>&nbsp;<\/p>\n<h5><span style=\"color: #008000;\"><strong>Reminder of my forecasting model for 2024<\/strong><\/span><\/h5>\n<p>I am using a top down model to predict total number of seats won by the major parties based on the sum and difference in the Conservative and Labour party average vote share per seat.<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-5731\" src=\"https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/07\/GE24-Forecast-GB-20240703.png\" alt=\"\" width=\"1578\" height=\"875\" srcset=\"https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/07\/GE24-Forecast-GB-20240703.png 1578w, https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/07\/GE24-Forecast-GB-20240703-300x166.png 300w, https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/07\/GE24-Forecast-GB-20240703-1024x568.png 1024w, https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/07\/GE24-Forecast-GB-20240703-768x426.png 768w, https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/07\/GE24-Forecast-GB-20240703-1536x852.png 1536w, https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/07\/GE24-Forecast-GB-20240703-450x250.png 450w, https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/07\/GE24-Forecast-GB-20240703-1320x732.png 1320w\" sizes=\"auto, (max-width: 1578px) 100vw, 1578px\" \/><\/p>\n<p>Once I&#8217;ve estimated sum seat share <strong><span style=\"color: #0000ff;\">CON<\/span>+<span style=\"color: #ff0000;\">LAB<\/span><\/strong> and seat share difference <strong><span style=\"color: #0000ff;\">CON<\/span>&#8211;<span style=\"color: #ff0000;\">LAB<\/span><\/strong>, I will have two simultaneous equations which can be solved to estimate the percentage of the <strong>632<\/strong> seats in Great Britain that will be won by the Conservative and Labour parties.\u00a0 <a href=\"https:\/\/marriott-stats.com\/nigels-blog\/uk-general-election-2024-forecasting-model-gb\/\" target=\"_blank\" rel=\"noopener\">I explain my model in full here<\/a>.\u00a0\u00a0For now, my model does not yet split seats won by other parties into the separate parties e.g. Lib Dems, SNP, etc.\u00a0 I am working on this and will include this feature when ready.<\/p>\n<p>One of the questions I have to resolve is which of the fitted lines in the above charts should I use.\u00a0 In fact I will be using simulation where each line has a certain probability of being chosen.<\/p>\n<p>For <strong><span style=\"color: #0000ff;\">CON<\/span><span style=\"color: #0000ff;\">+<\/span><span style=\"color: #ff0000;\">LAB<\/span><\/strong>, the probabilities I&#8217;ve decided to use are &#8211;<\/p>\n<ul>\n<li>Black line<strong> 1\/13<\/strong>\u00a0&#8211; it&#8217;s been over <strong>30<\/strong> years since this was the fit hence the low probability<\/li>\n<li>Green line <strong>9\/13<\/strong>\u00a0&#8211; the last <strong>7<\/strong> elections have used this fit hence why I say it is the most likely<\/li>\n<li>Brown line <strong>3\/13<\/strong>\u00a0&#8211; given the high probability of a new low for <strong><span style=\"color: #0000ff;\">CON<\/span><span style=\"color: #0000ff;\">+<\/span><span style=\"color: #ff0000;\">LAB<\/span><\/strong> vote share, I am not convinced the green line will hold hence why I&#8217;ve used a weight of 3 to 1 between the green and brown lines.<\/li>\n<li>Overall the three lines have probabilities in a <strong>1<\/strong> to <strong>3<\/strong> to <strong>9<\/strong> ratio.\u00a0 I like to use fixed ratios to avoid tweaking weights to get what I want!<\/li>\n<\/ul>\n<p>For <strong><span style=\"color: #0000ff;\">CON<\/span>&#8211;<span style=\"color: #ff0000;\">LAB<\/span><\/strong>, the probabilities I&#8217;ve decided to use are &#8211;<\/p>\n<ul>\n<li>Black line <strong>1\/4<\/strong>\u00a0&#8211; the last 3 elections have sat on this fit so it can&#8217;t be ruled out.<\/li>\n<li>Green line <strong>3\/4<\/strong> &#8211; the last 3 times where Labour kicked out a Conservative government with a big swing (<strong>1929, 1945, 1997<\/strong>) all sat on the green line.\u00a0 Polls are pointing to something similar in <strong>2024<\/strong>.<\/li>\n<li>Again I am using a <strong>1 to 3<\/strong> ratio between the two lines.<\/li>\n<\/ul>\n<p>The most likely outcome is that both models use their green fitted lines.\u00a0 This is what I will use to illustrate my forecast.<\/p>\n<p>&nbsp;<\/p>\n<h5><span style=\"color: #008000;\"><strong>Reminder of how accurate the polls are<\/strong><\/span><\/h5>\n<p>I recently realised <a href=\"https:\/\/marriott-stats.com\/nigels-blog\/uk-general-elections-how-accurate-are-opinion-polls-revised\/\" target=\"_blank\" rel=\"noopener\">voting intention polls on average do predict the vote share difference <strong><span style=\"color: #0000ff;\">CON<\/span>&#8211;<span style=\"color: #ff0000;\">LAB<\/span><\/strong> <\/a>provided one compares polls with average vote share per seat rather than national vote share.<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-5690\" src=\"https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/06\/PollErrors4-GE19-avg-vote-share-Diff.png\" alt=\"\" width=\"613\" height=\"301\" srcset=\"https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/06\/PollErrors4-GE19-avg-vote-share-Diff.png 613w, https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/06\/PollErrors4-GE19-avg-vote-share-Diff-300x147.png 300w, https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/06\/PollErrors4-GE19-avg-vote-share-Diff-450x221.png 450w\" sizes=\"auto, (max-width: 613px) 100vw, 613px\" \/><\/p>\n<p>When it comes to the sum vote share <strong><span style=\"color: #0000ff;\">CON<\/span>+<span style=\"color: #ff0000;\">LAB,<\/span><\/strong> polls are accurate on average over the whole timescale <strong>1950 to 2019<\/strong> but that hides notable differences over time.\u00a0 The last four elections, coinciding with a rise in number of pollsters and more web polling, underestimated the total vote share of the two main parties by <strong>3%<\/strong> on average.\u00a0 The step change in <strong>2010<\/strong> is statistically significant.<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-5691\" src=\"https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/06\/PollErrors4-GE19-avg-vote-share-Sum.png\" alt=\"\" width=\"613\" height=\"307\" srcset=\"https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/06\/PollErrors4-GE19-avg-vote-share-Sum.png 613w, https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/06\/PollErrors4-GE19-avg-vote-share-Sum-300x150.png 300w, https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/06\/PollErrors4-GE19-avg-vote-share-Sum-450x225.png 450w\" sizes=\"auto, (max-width: 613px) 100vw, 613px\" \/><\/p>\n<p>&nbsp;<\/p>\n<h5><span style=\"color: #008000;\"><strong>What do the latest polls say?<\/strong><\/span><\/h5>\n<p>For the <a href=\"https:\/\/x.com\/MarriottNigel\/status\/1808543302263976259\" target=\"_blank\" rel=\"noopener\">week ending <\/a><strong>3rd July 2024,<\/strong> using published data from <strong>16<\/strong> pollsters as of 2000 hours, <img loading=\"lazy\" decoding=\"async\" class=\" wp-image-5732 alignright\" src=\"https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/07\/GE24-Polls-Summary-20240703b.png\" alt=\"\" width=\"243\" height=\"253\" srcset=\"https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/07\/GE24-Polls-Summary-20240703b.png 334w, https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/07\/GE24-Polls-Summary-20240703b-288x300.png 288w\" sizes=\"auto, (max-width: 243px) 100vw, 243px\" \/>Labour has a <strong>17.8%<\/strong>\u00a0lead over the Conservatives.\u00a0 This would be a new record for Labour if this occurs on 4th July.<\/p>\n<p>At the same time, the combined <strong><span style=\"color: #0000ff;\">CON<\/span>+<span style=\"color: #ff0000;\">LAB<\/span><\/strong> vote share of <strong>60.6%<\/strong> would be the lowest on record since the <span style=\"color: #0000ff;\"><strong>CON<\/strong><\/span>\/<span style=\"color: #ff0000;\"><strong>LAB<\/strong><\/span> duopoly began in 1922, beating the <strong>66.6%<\/strong> in 2010.<\/p>\n<p>Reform are in 3rd place at <span style=\"color: #3366ff;\"><strong>16%<\/strong><\/span> and are holding steady.\u00a0 For reference, UKIP&#8217;s national vote share in 2015 was <span style=\"color: #3366ff;\"><strong>12.9%<\/strong><\/span>.<\/p>\n<p>The Lib Dems are in 4th place on <span style=\"color: #ff6600;\"><strong>11%<\/strong><\/span> and are holding steady.\u00a0 For reference, the Lib Dems national vote share in 2019 was <span style=\"color: #ff6600;\"><strong>11.8%<\/strong><\/span>.<\/p>\n<p>The Greens are in 5th place on <span style=\"color: #008000;\"><strong>6.5%<\/strong><\/span>.\u00a0 This would easily be their best ever general election beating their previous highest national vote share of <span style=\"color: #008000;\"><strong>3.8%<\/strong><\/span> in 2015.<\/p>\n<p>&nbsp;<\/p>\n<h5><span style=\"color: #008000;\"><strong>Will the polls get it wrong in 2024?<\/strong><\/span><\/h5>\n<p>First we need to define what we mean by &#8220;<em>wrong<\/em>&#8220;.<\/p>\n<p>I have usually used a difference between actual and polled vote share of <strong>4<\/strong> percentage points as the definition of a &#8220;<em>major error<\/em>&#8220;.\u00a0 That is because errors on this scale can cause <strong>30<\/strong> seats or more to change hands, often changing the narrative of the election.\u00a0 Conversely, any error less than <strong>2<\/strong> percentage points is not an error in my view.\u00a0 So I arrive at the following definition of &#8220;<em>wrong<\/em>&#8220;.<\/p>\n<ul>\n<li><strong>No Error<\/strong> &#8211; Errors less than <strong>2<\/strong> percentage points<\/li>\n<li><strong>Minor Error<\/strong> &#8211; Errors between <strong>2<\/strong> &amp; <strong>4<\/strong> percentage points<\/li>\n<li><strong>Major Error<\/strong> &#8211; Errors greater than <strong>4<\/strong> percentage points<\/li>\n<\/ul>\n<p>For the Conservative lead over Labour (<strong><span style=\"color: #0000ff;\">CON<\/span>&#8211;<span style=\"color: #ff0000;\">LAB<\/span><\/strong> vote share) statistic as estimated by the polls and compared to actual <strong><span style=\"color: #0000ff;\">CON<\/span>&#8211;<span style=\"color: #ff0000;\">LAB<\/span><\/strong> using average vote share per seat, there have been <strong>5<\/strong> major errors in the <strong>20<\/strong> elections since 1950.\u00a0 They are <span style=\"color: #ff0000;\"><strong>1950, 1983 &amp; 2017<\/strong> <\/span>where Labour did better than expected and <strong><span style=\"color: #0000ff;\">1992 &amp; 2015<\/span><\/strong> when the Conservatives did better than expected.\u00a0 Overall, <strong>25%<\/strong> of elections have experienced a major error for the Conservative lead over Labour.<\/p>\n<p>For the 2024 general election, I put the probability of a major error for the Conservative lead over Labour at greater than <strong>50%<\/strong>. \u00a0<img loading=\"lazy\" decoding=\"async\" class=\" wp-image-5725 alignright\" src=\"https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/07\/PollErrors7.png\" alt=\"\" width=\"394\" height=\"421\" srcset=\"https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/07\/PollErrors7.png 556w, https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/07\/PollErrors7-281x300.png 281w, https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/07\/PollErrors7-328x350.png 328w\" sizes=\"auto, (max-width: 394px) 100vw, 394px\" \/> There are two reasons why I put it this high.<\/p>\n<p>First is looking back at elections with a similar narrative to 2024.\u00a0 Throughout this election, the narrative has been &#8220;<em>How large will Labour&#8217;s majority be?<\/em>&#8221;\u00a0 The very first election I remember following was <strong>1983<\/strong> and that had an identical narrative except this time it was &#8220;<em>How large will the Tories majority be?<\/em>&#8221;\u00a0 I would identify <strong>1966, 1997, 2001<\/strong> and <strong>2017<\/strong> as similar elections.\u00a0 Those <strong>5<\/strong> elections are highlighted in the graphic here.<\/p>\n<p>In all <strong>5<\/strong> elections, the party that was expected to get hammered did better than the polls.\u00a0 The final lead for the party expected to get a large majority ended up on average being<strong> 3<\/strong> percentage points smaller than expected.\u00a0 Currently Labour have a <span style=\"color: #ff0000;\"><strong>18%<\/strong><\/span> lead in the polls over the Conservatives but if my analysis of history is right, it may end up at <span style=\"color: #ff0000;\"><strong>15%<\/strong><\/span>\u00a0instead.<\/p>\n<p>The second reason is the pollsters can be divided into two camps based on whether they attempt to squeeze or model those saying don&#8217;t know to the pollsters.\u00a0 A\u00a0 squeeze question is something like &#8220;S<em>uppose you were required to vote by law (as in Australia), which party would you vote for?<\/em>&#8221;\u00a0 Based on <a href=\"https:\/\/x.com\/jeffgrim71\/status\/1808002591537713155\" target=\"_blank\" rel=\"noopener\">this list from X\/Twitter<\/a>, it appears half of the pollsters are Squeezers and half just ignore those saying Don&#8217;t Know.<\/p>\n<p>On average <a href=\"https:\/\/x.com\/MarriottNigel\/status\/1808543311084327301\" target=\"_blank\" rel=\"noopener\">the Squeezers have <span style=\"color: #0000ff;\"><strong>CON<\/strong><\/span>&#8211;<span style=\"color: #ff0000;\"><strong>LAB<\/strong><\/span> leads <strong>1%<\/strong> smaller than non-Squeezers<\/a>.\u00a0 That combined with my first observation from elections with similar narratives to 2024 is why I think a major error on the <span style=\"color: #0000ff;\"><strong>CON<\/strong><\/span>&#8211;<span style=\"color: #ff0000;\"><strong>LAB<\/strong><\/span> lead is more likely than not.<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone wp-image-5734\" src=\"https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/07\/GE24-Polls-Summary-20240703-Squeeze.png\" alt=\"\" width=\"495\" height=\"173\" srcset=\"https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/07\/GE24-Polls-Summary-20240703-Squeeze.png 592w, https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/07\/GE24-Polls-Summary-20240703-Squeeze-300x105.png 300w, https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/07\/GE24-Polls-Summary-20240703-Squeeze-450x157.png 450w\" sizes=\"auto, (max-width: 495px) 100vw, 495px\" \/><\/p>\n<p>&nbsp;<\/p>\n<h5><span style=\"color: #008000;\"><strong>How do I use simulation to make my forecast?<\/strong><\/span><\/h5>\n<p>My official forecast is an average of <strong>10,000<\/strong> simulations of possible scenarios.\u00a0 The process works like this.<\/p>\n<p>My first simulation assumes the polls are correct and that both the sum seat share and seat share difference models use the green fit lines.\u00a0 I call this the <strong>Illustrative Example<\/strong> and the calculation works like this.<\/p>\n<ul>\n<li>The estimated sum seat share<strong><span style=\"color: #0000ff;\"> CON<\/span>+<span style=\"color: #ff0000;\">LAB<\/span><\/strong> is <strong>87.6% <\/strong>using the green fit line\u00a0i.e. Other parties will win <strong>12.4%<\/strong> of the <strong>632<\/strong> seats in Great Britain.<\/li>\n<li>The estimated seat share difference <strong><span style=\"color: #0000ff;\">CON<\/span>&#8211;<span style=\"color: #ff0000;\">LAB<\/span><\/strong> is <span style=\"color: #ff0000;\"><strong>-48.0%<\/strong><span style=\"color: #000000;\"> using the green fit line<\/span>.<\/span><\/li>\n<li>The estimated <span style=\"color: #0000ff;\"><strong>CON<\/strong><\/span> seat share is <span style=\"color: #0000ff;\"><strong>19.8%<\/strong><\/span> = (sum+diff)\/2 = ( <strong>87.6%<\/strong> + <span style=\"color: #ff0000;\"><strong>-48%<\/strong><\/span> )\/2<\/li>\n<li>The estimated <span style=\"color: #ff0000;\"><strong>LAB<\/strong><\/span> seat share is <span style=\"color: #ff0000;\"><strong>67.8%<\/strong><\/span> = (sum-diff)\/2 = ( <strong>87.6%<\/strong> &#8211; <span style=\"color: #ff0000;\"><strong>-48%<\/strong><\/span> )\/2<\/li>\n<\/ul>\n<p>With<strong> 632<\/strong> seats in all of Great Britain, the illustrative example results in <span style=\"color: #0000ff;\"><strong>125 CON<\/strong><\/span>, <span style=\"color: #ff0000;\"><strong>429 LAB<\/strong><\/span> and <strong>78 OTH<\/strong> seats.<\/p>\n<p>Thereafter, the other <strong>9,999<\/strong> simulations are processed as follows &#8211;<\/p>\n<ol>\n<li>A random number is added to the <strong><span style=\"color: #0000ff;\">CON<\/span>+<span style=\"color: #ff0000;\">LAB<\/span><\/strong> vote share sum indicated by the polls.\u00a0 Given the notable step change in 2010, I am currently drawing a random number from a normal distribution with mean +<strong>2.0%<\/strong> and standard deviation <strong>1.0%<\/strong>.\u00a0 I am keeping this element under review.<\/li>\n<li>A random number between <strong>0<\/strong> and <strong>1<\/strong> is created and this decides whether the fitted model for sum seat share <strong><span style=\"color: #0000ff;\">CON<\/span>+<span style=\"color: #ff0000;\">LAB<\/span><\/strong> uses the black (<strong>8%<\/strong>), green (<strong>69%<\/strong>) or brown(<strong>23%<\/strong>) fitted lines.<\/li>\n<li>The chosen fitted line is used to estimate the sum seat share <strong><span style=\"color: #0000ff;\">CON<\/span>+<span style=\"color: #ff0000;\">LAB<\/span><\/strong>.<\/li>\n<li>To the estimated sum seat share <strong><span style=\"color: #0000ff;\">CON<\/span>+<span style=\"color: #ff0000;\">LAB<\/span><\/strong>, a random number drawn from a normal distribution with mean <strong>0%<\/strong> and standard deviation <strong>0.5%<\/strong> is added to allow for residual error.<\/li>\n<li>A random number is added to the <strong><span style=\"color: #0000ff;\">CON<\/span>&#8211;<span style=\"color: #ff0000;\">LAB<\/span><\/strong> vote share difference indicated by the polls.\u00a0 This is more straightforward as it is a random draw from a normal distribution with mean <strong>0.0%<\/strong> and standard deviation of <strong>2.5%<\/strong>.\n<ul>\n<li><span style=\"color: #993300;\"><em>Important &#8211; this step does not include the expected major error I refer to above.\u00a0 If I were, I would change the mean to <strong>+2%<\/strong> or <strong>+3%<\/strong>.<\/em><\/span><\/li>\n<\/ul>\n<\/li>\n<li>A random number between <strong>0<\/strong> and <strong>1<\/strong> is created and this decides whether the fitted model for seat share difference <strong><span style=\"color: #0000ff;\">CON<\/span>&#8211;<span style=\"color: #ff0000;\">LAB <\/span><\/strong>uses the black (<strong>25%)<\/strong> or green (<strong>75%<\/strong>) fitted lines.<\/li>\n<li>The chosen fitted line is used to estimate the seat share difference <strong><span style=\"color: #0000ff;\">CON<\/span>&#8211;<span style=\"color: #ff0000;\">LAB<\/span><\/strong>.<\/li>\n<li>To the estimated seat share difference <strong><span style=\"color: #0000ff;\">CON<\/span>&#8211;<span style=\"color: #ff0000;\">LAB<\/span><\/strong>, a random number drawn from a normal distribution with mean <strong>0%<\/strong> and standard deviation <strong>2.5%<\/strong> is added to allow for residual error.<\/li>\n<\/ol>\n<p>On completion of all simulations, the average of the estimated sum seat share <strong><span style=\"color: #0000ff;\">CON<\/span><span style=\"color: #0000ff;\">+<\/span><span style=\"color: #ff0000;\">LAB <\/span><\/strong>after <img loading=\"lazy\" decoding=\"async\" class=\" wp-image-5735 alignright\" src=\"https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/07\/GE24-Forecast-GB-20240703-Seats.png\" alt=\"\" width=\"340\" height=\"352\" srcset=\"https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/07\/GE24-Forecast-GB-20240703-Seats.png 493w, https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/07\/GE24-Forecast-GB-20240703-Seats-290x300.png 290w, https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/07\/GE24-Forecast-GB-20240703-Seats-338x350.png 338w\" sizes=\"auto, (max-width: 340px) 100vw, 340px\" \/>step 4 and the estimated seat share difference <strong><span style=\"color: #0000ff;\">CON<\/span>&#8211;<span style=\"color: #ff0000;\">LAB <\/span><\/strong>after step 8 are calculated.\u00a0 These are then solved to get the estimated number of seats to be won by the Conservatives, Labour and Others.\u00a0 The <strong>10,000<\/strong> simulations provide data on possible margins of error and are summarised in the table below.\u00a0 An explanation of the table <a href=\"https:\/\/x.com\/MarriottNigel\/status\/1807745075339784478\" target=\"_blank\" rel=\"noopener\">is given in this X\/Twitter thread<\/a> as well.<\/p>\n<p>&nbsp;<\/p>\n<p><strong>How do I forecast the smaller parties?<\/strong><\/p>\n<p>The table at the end of the previous section summarises my simulations and only gives estimates for the Conservatives, Labour and Others across Great Britain.\u00a0 <a href=\"https:\/\/marriott-stats.com\/nigels-blog\/uk-general-election-2024-forecasting-model-gb\/\" target=\"_blank\" rel=\"noopener\">In my article explaining my model in detail<\/a>, I stated I would run a similar forecasting exercise for England only for the following reasons &#8211;<\/p>\n<ol>\n<li>The expected number of seats for Other parties in England will be mostly Liberal Democrat wins.<\/li>\n<li>The difference between expected number of seats for Others in Great Britain and in England will capture mostly SNP and Plaid Cymru wins.<\/li>\n<\/ol>\n<p>Any further splits into wins for Greens, Reform and Liberal Democrats in Scotland &amp; Wales will be a finger in the air job by me based mostly on what I have read from other election analysts on X\/Twitter.\u00a0 I have not had the time to do detailed seat modelling for this election so as far as I am concerned, I should be judged on how many seats are won by the Conservatives and Labour.<\/p>\n<p>The forecasting process is exactly the same as for Great Britain, the only differences are the polls are based on England only crossbreaks of the national polls and the models for <strong><span style=\"color: #0000ff;\">CON<\/span>+<span style=\"color: #ff0000;\">LAB<\/span><\/strong> and <span style=\"color: #0000ff;\"><strong>CON<\/strong><\/span>&#8211;<span style=\"color: #ff0000;\"><strong>LAB<\/strong><\/span> are the ones shown in the chart below.<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-5724\" src=\"https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/07\/GE24-Forecast-ENG-20240703.png\" alt=\"\" width=\"1578\" height=\"875\" srcset=\"https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/07\/GE24-Forecast-ENG-20240703.png 1578w, https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/07\/GE24-Forecast-ENG-20240703-300x166.png 300w, https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/07\/GE24-Forecast-ENG-20240703-1024x568.png 1024w, https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/07\/GE24-Forecast-ENG-20240703-768x426.png 768w, https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/07\/GE24-Forecast-ENG-20240703-1536x852.png 1536w, https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/07\/GE24-Forecast-ENG-20240703-450x250.png 450w, https:\/\/marriott-stats.com\/nigels-blog\/wp-content\/uploads\/2024\/07\/GE24-Forecast-ENG-20240703-1320x732.png 1320w\" sizes=\"auto, (max-width: 1578px) 100vw, 1578px\" \/><\/p>\n<p>After running <strong>10,000<\/strong> simulations for England, the outcome was the forecast <strong>82<\/strong>\u00a0seats for Others in Great Britain split into <strong>53<\/strong> for England and <strong>29<\/strong>\u00a0for Scotland and Wales.<\/p>\n<p>Now comes the finger in the air bit!\u00a0 I&#8217;ve decided to go with the following round numbers &#8211;<\/p>\n<ul>\n<li><strong>LD<\/strong> &#8211; <strong>50<\/strong> seats<\/li>\n<li><strong>SNP<\/strong> &#8211; <strong>25<\/strong> seats<\/li>\n<li><strong>PC<\/strong> &#8211; <strong>3<\/strong> seats<\/li>\n<li><strong>GRN<\/strong> &#8211; <strong>2<\/strong> seats<\/li>\n<li><strong>REF<\/strong> &#8211; <strong>3<\/strong> seats<\/li>\n<\/ul>\n<p>It&#8217;s worth making the following points about the Liberal Democrats because I see a lot of excitement on social media.<\/p>\n<ul>\n<li>Their national vote share in 2019 was <strong>11.8%<\/strong> &amp; latest polls suggest they will drop to <strong>11%<\/strong> in 2024.<\/li>\n<li>In 1997, the LD national vote share fell to <strong>17.2%<\/strong> from <strong>18.3%<\/strong> in 1992 but the number of seats they won rose from <strong>20<\/strong> to <strong>46<\/strong>.<\/li>\n<li>For 2024, I am predicting they will go from <strong>11<\/strong> to <strong>50<\/strong> seats but starting at a lower base and falling back a bit in terms of votes.<\/li>\n<li>The main reason is the Conservatives are expected to do worse than 1997.<\/li>\n<\/ul>\n<p>A final point to note is anti-Conservative tactical voting.\u00a0 This was a recognised factor in 1997 and some people think we will see the same effect in 2024.\u00a0 The problem with that argument is the Lib Dem vote of 2019 already includes a lot of tactical voting built up over 2017 &amp; 2019 due to remain voters seeking to frustrate Brexit.\u00a0 Therefore I ask how much more tactical voting can we expect in 2024?<\/p>\n<p>Putting all of this together, I would be surprised if the Lib Dems got more than <strong>50<\/strong> seats.\u00a0 I seriously thought about reducing the forecast to <strong>45<\/strong> seats but I wasn&#8217;t sure where to put the other <strong>5<\/strong> seats.\u00a0 That is why I have stuck with <strong>50<\/strong>.<\/p>\n<p>&nbsp;<\/p>\n<h5><span style=\"color: #008000;\"><strong>My Official Forecast for the 2024 UK General Election<\/strong><\/span><\/h5>\n<p>As of <strong>3rd July 2024<\/strong>, my simulations gave me my final official forecast as follows &#8211;<\/p>\n<ul>\n<li><span style=\"color: #0000ff;\"><strong>CON 130<\/strong>\u00a0<\/span>seats (<span style=\"color: #0000ff;\">113 to 150<\/span>)<\/li>\n<li><strong><span style=\"color: #ff0000;\">LAB 420<\/span><\/strong>\u00a0seats (<span style=\"color: #ff0000;\">436 to 408<\/span>)<\/li>\n<li><span style=\"color: #ff6600;\"><strong>LD 50<\/strong><\/span><\/li>\n<li><strong>SNP 25<\/strong><\/li>\n<li><span style=\"color: #008000;\"><strong>GRN 3<\/strong><\/span><\/li>\n<li><span style=\"color: #339966;\"><strong>PC 2<\/strong><\/span><\/li>\n<li><span style=\"color: #00ccff;\"><strong>REF 2<\/strong><\/span><\/li>\n<li><strong>OTH NI 18<\/strong><\/li>\n<\/ul>\n<p>The ranges in brackets represent the interquartile range of my <strong>10,000<\/strong> scenarios i.e. from the <strong>25th<\/strong> to the <strong>75th<\/strong> percentile.\u00a0 Note the error in <span style=\"color: #0000ff;\"><strong>CON<\/strong><\/span> and <span style=\"color: #ff0000;\"><strong>LAB<\/strong><\/span> are inversely correlated hence why Labour&#8217;s range is from high to low i.e. if the Conservatives do end up with <span style=\"color: #0000ff;\"><strong>113<\/strong><\/span>\u00a0seats, Labour are more likely to get around <strong><span style=\"color: #ff0000;\">436<\/span>\u00a0<\/strong>seats.<\/p>\n<p>Note this forecast does not include a major polling error in the Conservative lead over Labour as discussed earlier.\u00a0 If that were to happen, the Conservatives would be <span style=\"color: #0000ff;\"><strong>10<\/strong><\/span> seats higher and Labour <strong><span style=\"color: #ff0000;\">10<\/span><\/strong> seats lower.<\/p>\n<p>Compared to <a href=\"https:\/\/marriott-stats.com\/nigels-blog\/uk-general-election-2024-forecast-2\/\" target=\"_blank\" rel=\"noopener\">my forecast of 30th June<\/a>, there has been a shift of <strong>17<\/strong> expected seats from Labour to the Conservatives.\u00a0 This is because the polls show Labour&#8217;s lead over the Conservatives has narrowed.<\/p>\n<p>&nbsp;<\/p>\n<h5><em><strong><span style=\"color: #993300;\">Postscript &#8211;<\/span><\/strong><\/em><\/h5>\n<p><em><span style=\"color: #993300;\">Should there be any significant updates to the polls, I will state my revised forecast here.<\/span><\/em><\/p>\n<p>&nbsp;<\/p>\n<p><strong><span style=\"color: #993300;\">&#8212; Would you like to comment on this article? &#8212;-<\/span><\/strong><\/p>\n<p>Please do leave your comments on this <span style=\"color: #008000;\"><a style=\"color: #008000;\" href=\"https:\/\/x.com\/MarriottNigel\/status\/1808592941608313336\" target=\"_blank\" rel=\"noopener\"><strong>X\/Twitter<\/strong> <\/a><\/span>thread.<\/p>\n<h5><strong><span style=\"color: #993300;\">&#8212; Subscribe to my newsletter to receive more articles like this one! &#8212;-<\/span><\/strong><\/h5>\n<p>If you would like to receive notifications from me of news, articles and offers relating to Elections &amp; Polling, please <span style=\"color: #008000;\"><strong><a style=\"color: #008000;\" href=\"https:\/\/marriott-stats.com\/nigels-blog\/subscribe-to-our-newsletter\/\" target=\"_blank\" rel=\"noopener\">click here to go to my Newsletter Subscription page<\/a><\/strong><\/span> and tick the Elections and\/or Surveys category and other categories that may be of interest to you.\u00a0 You will be able to unsubscribe at anytime.<\/p>\n<p>More articles on elections can be found by clicking on the <a href=\"https:\/\/marriott-stats.com\/nigels-blog\/category\/elections\/\" target=\"_blank\" rel=\"noopener\">Elections<\/a> tab at the top of your screen.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>My third and final forecast of the number of seats to be won by each party in the 2024 UK General Election is &#8211; CON 130 (+17) LAB 420 (-17) LD 50 SNP 25 GRN 3 PC 2 REF 2 OTH NI 18 The numbers in brackets are the changes from my second forecast of [&hellip;]<\/p>\n","protected":false},"author":3,"featured_media":5735,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_genesis_hide_title":false,"_genesis_hide_breadcrumbs":false,"_genesis_hide_singular_image":false,"_genesis_hide_footer_widgets":false,"_genesis_custom_body_class":"","_genesis_custom_post_class":"","_genesis_layout":"","footnotes":""},"categories":[2,6],"tags":[21,19,25,215,45,333],"class_list":{"0":"post-5722","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"has-post-thumbnail","7":"category-elections","8":"category-forecasting","9":"tag-election-forecasting","10":"tag-elections","11":"tag-forecasting-model","12":"tag-general-election-2024","13":"tag-seat-forecast","14":"tag-seats","15":"entry","16":"override"},"_links":{"self":[{"href":"https:\/\/marriott-stats.com\/nigels-blog\/wp-json\/wp\/v2\/posts\/5722","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/marriott-stats.com\/nigels-blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/marriott-stats.com\/nigels-blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/marriott-stats.com\/nigels-blog\/wp-json\/wp\/v2\/users\/3"}],"replies":[{"embeddable":true,"href":"https:\/\/marriott-stats.com\/nigels-blog\/wp-json\/wp\/v2\/comments?post=5722"}],"version-history":[{"count":10,"href":"https:\/\/marriott-stats.com\/nigels-blog\/wp-json\/wp\/v2\/posts\/5722\/revisions"}],"predecessor-version":[{"id":5741,"href":"https:\/\/marriott-stats.com\/nigels-blog\/wp-json\/wp\/v2\/posts\/5722\/revisions\/5741"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/marriott-stats.com\/nigels-blog\/wp-json\/wp\/v2\/media\/5735"}],"wp:attachment":[{"href":"https:\/\/marriott-stats.com\/nigels-blog\/wp-json\/wp\/v2\/media?parent=5722"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/marriott-stats.com\/nigels-blog\/wp-json\/wp\/v2\/categories?post=5722"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/marriott-stats.com\/nigels-blog\/wp-json\/wp\/v2\/tags?post=5722"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}