Q4 AI Forecasting Benchmark Tournament practice run of bestworldbot 10-9-2024

Who will be the next US President?

Botmaster Jeremy’s bestworldbot, competing in the Q4 AI Forecasting Benchmark Tournament for a prize pool of $30,000, practiced on four questions today.

The question “Will Joe Biden be elected US President in 2024?” has an obvious answer: no.  This question was helpful for testing an AI bot because we already know the answer: bestworldbot aced this one.

Similarly, we already know that polls show that the race between Kamala Harris and Donald Trump are close: bestworldbot is showing Trump with a slight edge.

Botmaster Jeremy’s team includes BestWorld’s Brian LaBatte, Michał Dubrawski, Bruce Lawhorn, and Carolyn Meinel.

Click here for the winners of the Q3 AI Forecasting Benchmark Tournament.

Metaculus’ analyses of the competition’s bots’ capabilities.

The final leaderboard for the Q3 AI Forecasting Benchmark Tournament.

Our Humans vs Multi-AI Panel Experiment of July 25 — Sept. 18, 2024 competition.

All of our bots vs humans experiments.

Our concluded FED rate increase competition.

All of BestWorld’s current projects.


















Will Kamala Harris be elected US President in 2024?

Obtaining outside newsfeed.
Summarizing outside newsfeed.
Querying Perplexity (AI predicts: 0.52 – confidence: 6)
Querying Claude (AI predicts: 0.48 – confidence: 5)
Querying Mistral (AI predicts: 0.4 – confidence: 6)
Querying Cohere (AI predicts: 0.48 – confidence: 7)
Querying OpenAI (AI predicts: 0.48 – confidence: 6)

Median from LLMs: 0.48

Base rate: 0.48 (from OpenAI)

SD: 0.04
Confidence: 6
Conf Mode: Normal
Mellers: 0.46
Reverse Mellers: 0.49
Theory of Mind: 0.5 (What did the LLMs think?)
Beta Distribution: 0.5
# LLM responses: 5
Submitted value: 0.48

We queried Perplexity, Claude, Mistral, Cohere, OpenAI.
A summary of what the responding AIs considered is as follows:

The current political landscape for the 2024 US presidential election showcases a tightly contested race between Kamala Harris and Donald Trump. Harris holds a slight lead in national polls, with approximately a 3-point advantage, yet betting markets give Trump a marginal edge. The outcome of the election heavily relies on key swing states like Arizona, Georgia, Pennsylvania, and Wisconsin, where both candidates have significant competition. Factors such as Harris’s debate performance, the timing of Biden’s withdrawal, and the historical complexities of election prediction contribute to the overall uncertainty. Additionally, variables such as potential polling inaccuracies, shifts in voter sentiment, and the impact of the Electoral College could sway the results unexpectedly, making the election’s outcome far from certain.
Submitting to Metaculus.

Will Donald Trump be elected US President in 2024?

Obtaining outside newsfeed.
Summarizing outside newsfeed.
Querying Perplexity (AI predicts: 0.51 – confidence: 6)
Querying Claude (AI predicts: 0.53 – confidence: 5)
Querying Mistral (AI predicts: 0.55 – confidence: 6)
Querying Cohere (AI predicts: 0.51 – confidence: 6)
Querying OpenAI (AI predicts: 0.51 – confidence: 6)

Median from LLMs: 0.51

Base rate: 0.51 (from OpenAI)

SD: 0.02
Confidence: 6
Conf Mode: Normal
Mellers: 0.52
Reverse Mellers: 0.51
Theory of Mind: 0.5 (What did the LLMs think?)
Beta Distribution: 0.5
# LLM responses: 5
Submitted value: 0.51

We queried Perplexity, Claude, Mistral, Cohere, OpenAI.
A summary of what the responding AIs considered is as follows:

The predictions about the 2024 U.S. presidential race between Donald Trump and Kamala Harris primarily rest on polling data and betting markets, which suggest Trump has a slight edge. Trump is ahead in key battleground states such as Arizona, Georgia, and North Carolina, but the margins are tight and within the margin of error, indicating a highly competitive race. The outcome is uncertain and will be influenced by factors such as voter turnout and potential late shifts in the political landscape. However, the prediction may be flawed due to the volatility inherent in elections, the potential for last-minute shifts in voter sentiment, and the reliability of external newsfeeds. Unforeseen events and campaign strategies in the remaining weeks could significantly alter the result, and history has shown polling data can sometimes be misleading. Lastly, the influence of current President Joe Biden’s decisions could also play a role in the final outcome.
Submitting to Metaculus.

Will Joe Biden be elected US President in 2024?

Obtaining outside newsfeed.
Summarizing outside newsfeed.

Querying Perplexity (AI predicts: 0 – confidence: 10)
Querying Claude (AI predicts: 0 – confidence: 10)
Querying Mistral (AI predicts: 0.05 – confidence: 8)
Querying Cohere(Error: Cohere is not returning an array)
Querying OpenAI (AI predicts: 0 – confidence: 10)

Median from LLMs: 0

Base rate: 0 (from OpenAI)

SD: 0.02
Confidence: 10
Conf Mode: High
Mellers: 0
Reverse Mellers: 0
Theory of Mind: 0 (What did the LLMs think?)
Beta Distribution: N/A
# LLM responses: 4
Submitted value: 0.001

We queried Perplexity, Claude, Mistral, OpenAI.
A summary of what the responding AIs considered is as follows:

Due to President Joe Biden’s official withdrawal from the U.S. presidential race on October 8, 2024, it is certain that he will not be re-elected as President in the 2024 elections. This decision has positioned Vice President Kamala Harris as the Democratic candidate, who will contend against former President Donald Trump. The election, taking place on November 5, 2024, is now a competition between Harris and Trump with no chance for Biden’s re-election. Although theoretically possible, scenarios like Biden’s return to the race or contested election results are considered extremely unlikely given the official confirmation of his withdrawal and the proximity to the election date.
Submitting to Metaculus.

Will the Real-time Sahm Rule Recession Indicator increase Year-over-Year for Nov 2024?

Obtaining outside newsfeed.
Summarizing outside newsfeed.
Querying Perplexity (AI predicts: 0.7 – confidence: 7)
Querying Claude (AI predicts: 0.65 – confidence: 6)
Querying Mistral (AI predicts: 0.7 – confidence: 6)
Querying Cohere (AI predicts: 0.5 – confidence: 5)
Querying OpenAI (AI predicts: 0.75 – confidence: 6)

Median from LLMs: 0.7

Base rate: 0.75 (from OpenAI) SD: 0.09
Confidence: 6
Conf Mode: Normal
Mellers: 0.82
Reverse Mellers: 0.6
Theory of Mind: 0.5 (What did the LLMs think?)
Beta Distribution: 0.75
# LLM responses: 5
Submitted value: 0.7

We queried Perplexity, Claude, Mistral, Cohere, OpenAI.
A summary of what the responding AIs considered is as follows:

The Sahm Rule Recession Indicator, which evaluates economic downturns based on the increase in unemployment rates, crossed the critical 0.50% threshold in August 2024, reaching 0.57%. This indicator is based on the three-month moving average of the national unemployment rate and signals a recession when it rises significantly relative to the prior year. Historically, once activated, this indicator tends to signal persistently high levels of unemployment for several months, suggesting potential economic challenges continuing into November 2024. However, there are uncertainties regarding the interpretation of an increase in unemployment as necessarily signaling a recession, as it might be affected by structural changes in the labor market or new labor market entrants rather than declining demand. Moreover, significant economic improvements or policy interventions could alter this trajectory before November, potentially stabilizing the unemployment rate and prompting downward revisions to the indicator based on new data.
Submitting to Metaculus.

Author