Current Forecast
| Harris | Trump |
|---|---|
| 267 | 271 |
This is our latest Forecast for the 2024 US General Election. Forecast is calculated on a “Nowcast” basis meaning that it is the result if an election were held today.
| Probability Trump Victory | 51.6% |
| Probability Trump 260+ EC Votes | 78.0% |
| Probability Trump 250+ EC Votes | 90.0% |
| Probability Trump 240+ EC Votes | 96.7% |
| Probability Tie | 0.4% |
| 50% Confidence Interval Trump EC Votes | 261 – 283 |
| Trump Modal EC Vote (EC Vote value with highest prob) | 277 – 7.8% likely |
| Trump Median EC Vote (Centre of Probability Distribution) | 271 – 3.6% likely |
State Forecasts – Prob Trump Winning
| State | Probability Trump Wins | Change Since Last Forecast |
| Arizona | 85.4% | Slight Improvement for Trump |
| Georgia | 64.5% | Strong Improvement for Harris |
| Michigan | 50.0% | Improvement for Trump |
| Nevada | 52.3% | Improvement for Harris |
| North Carolina | 89.0% | Strong Improvement for Trump |
| Pennsylvania | 23.0% | Very Strong Improvement for Harris |
| Wisconsin | 37.1% | Improvement for Trump |
Commentary
Our forecast is dominated by the collapse of Trump in polling in Pennsylvania after the remarks made in the Madison Square Gardens rally. The State has gone from being almost certain for him to having less than a 1 in 4 chance of taking it. Some improvement for Trump in Michigan and Wisconsin has partly mitigated that dramatic turnaround, but the election is now pretty much a coin toss.
There is a severe shortage of polling in Florida and we still can’t see any effect there of the remarks made during the rally.
We will provide daily updates moving forward.
Presidential Election Forecast Philosophy
Our US Presidential Election forecast is based on polling in individual states. We run a monte-carlo simulation across the entire country and then calculate the probability of each candidate winning each state or electoral district (in the case of Maine and Nebraska). From this we allocate Electoral College votes State by State giving a final total for each candidate.
Do you allocate whole States / Districts to Candidates?
We run 10,000 simulations of the election. Our seat forecast is based on allocating the total EC votes for each State / District to the candidate who “wins” in each simulation. From this we produce a probability distribution of likely outcomes across all the simulations in aggregate.
Some precition sites like FiveThirtyEight or Nate Silver have a different probability of Trump winning. What are you doing wrong?
Nothing! We all have slightly different modelling philosophies which produce different outcomes.
Can you tell me what the probability of a certain event happening is?
Sure. We provide the most interesting probabilities above, but leave a comment below if there is a particular probability you want us to share with you.
Peter, I wonder what you make of this Guardian article here which rather echoes my own gut feeling? “Dead-heat poll results are astonishing – and improbable, these experts say” see –
https://www.theguardian.com/us-news/2024/nov/02/what-polls-mean-so-far-trump-harris-election-voters
We’ll know on Wednesday morning!
Seen a lot about increase in early voting, do you anticipate this having any impact on likely outcome?
I haven’t seen much good evidence this favours either candidate