MBFC’s Last Election Sim Shows Harris with a 66.78% Win Rate

(Update) The 2024 Presidential Election is over, with Donald Trump winning all swing states. See the map above. While the Sim did not get the final outcome right, it did predict that Donald Trump would win 1661 times out of 5000. This sim was based primarily on polling averages and historical voting. The simulation exposed that the pollsters once again underestimated Donald Trump. However, the final AtlasIntel polls were very accurate. The sim also exposed that the Gender Gap was not a factor in this election. All in all, the sim was a success, considering it was a first run. The new data will help us sharpen it for the next elections in 2026 and 2028.

Media Bias Fact Check (MBFC) initially developed its Electoral College Simulation in 2023 to anticipate a possible Trump-Biden rematch. The algorithm was calibrated to simulate the 2020 election results accurately. The model accounted for a range of metrics, including historical results, polling data (state, national, congressional), fundraising, favorability, and margin of error factors (See Methodology).

In June 2024, the simulation showed Trump defeating Biden in up to 90+% of 5000 daily simulations. However, with Kamala Harris as the Democratic candidate, the model shifted, erasing Trump’s significant winning odds and ultimately positioning Harris as the favorite.

In our latest (final) simulation of 5,000 elections, Harris won the election 66.78% of the time, averaging 276 electoral votes to Trump’s 262. Analyzing the most common outcomes, Harris most frequently secured wins in the swing states of Pennsylvania, Nevada, Michigan, and Wisconsin, resulting in a 276-262 victory. The next most likely outcome showed Trump taking Pennsylvania, flipping the result to a 281-257 win in his favor. Less common results in order of likelihood include:

Harris 308 – 230: Harris wins Michigan, Wisconsin, Pennsylvania, Nevada, North Carolina, and Georgia. (12%)
Harris 292 – 246: Harris wins Michigan, Wisconsin, Pennsylvania, Nevada, and North Carolina. (8%)
Harris 270 – 268: Harris narrowly wins with Michigan, Wisconsin, and Pennsylvania, without Nevada. (7%)
Harris 319 – 219: Harris sweeps all the swing states. (3%)

While the model favors a Harris victory, MBFC does not speculate on specific outcomes, noting that polling can often misrepresent trends. For example, polls underestimated Trump’s support in the 2016 and 2020 elections, often by margins beyond the margin of error. Are they still underestimating him or overcompensating?

In this election, polling data reflects a notable gender gap, with Harris leading among women by as much as 20 points. This dynamic, if accurate, could influence voter turnout in Harris’s favor, potentially surpassing expectations set by previous election polls. While MBFC cannot predict the final outcome, the simulation provides a probability-based assessment that Harris will reach at least 270 electoral votes. Although the outcome remains uncertain, the final simulation indicates Harris’s advantage in the path to 270, particularly considering demographic factors like a significant gender gap. (See the Interactive Map)