The hottest Forecasting Substack posts right now

And their main takeaways
Category
Top Business Topics

It's 2026 and life is still messy

Klement on Investing • 4 implied HN points • 02 Jan 26
💰 Finance Forecasting
  1. Prefer experimental and empirical evidence over abstract economic theory, because investing should be based on how the world actually behaves.
  2. Markets and economies are messy, complex social systems with many second‑ and third‑round effects, so simple “ceteris paribus” models and daily market noise often mislead.
  3. Use a clear rhythm for thinking: focus on ESG, deep economic and finance topics midweek, and lighter, quirky economic takes on Fridays, while staying curious and ready to challenge assumptions.

Superforecasting the premises in “Is power-seeking AI an existential risk?”

Joe Carlsmith's Substack • 58 implied HN points • 18 Oct 23
🕹 Technology Forecasting
  1. Good Judgment solicited reviews and forecasts from superforecasters on the argument for AI risk.
  2. Superforecasters placed higher probabilities on some AI risk premises and lower on others compared to the original report.
  3. Author is skeptical of heavy updates based solely on superforecaster numbers and emphasizes the importance of object-level arguments.
Get a weekly roundup of the best Substack posts, by hacker news affinity:

Ensemble Weather Forecast API

Open-Meteo • 351 implied HN points • 05 Jun 23
🕹 Technology Forecasting
  1. Ensemble weather forecasts show a range of possibilities, helping to understand the uncertainty in predictions.
  2. Weather forecasts differ in reliability based on location and weather patterns, affecting the level of uncertainty in predictions.
  3. The Ensemble API combines various weather models, providing access to different weather variables for various purposes.

Buying Time

Comment is Freed • 93 implied HN points • 31 Oct 24
💰 Finance Forecasting
  1. The government is trying to slow down a financial crisis by making tough choices, but this is just a temporary fix. They've had to reverse some tax cuts and allow for more borrowing.
  2. The previous government's budget projections were unrealistic, leading to anger and concerns about how much spending was really needed. Many departments are facing serious issues.
  3. Despite these reversals, the government doesn't have much time left to make the situation better. They're in a tough spot and need to find new ways to improve the economy.

Lawler: Early Read on Existing Home Sales in December

CalculatedRisk Newsletter • 47 implied HN points • 17 Jan 25
💰 Finance Forecasting
  1. Existing home sales stayed steady at an annual rate of 4.15 million in December, the same as November. This shows slight improvement compared to the previous year.
  2. The average price of existing single-family homes increased by about 5.6% compared to last year. This indicates that home values are generally rising.
  3. The upcoming report from the National Association of Realtors is expected to show even higher sales this month. If confirmed, it would be a third month of increasing year-over-year sales.

Tang Shiping predicts Trump "highly likely" to win - probability more than 60%

Pekingnology • 56 implied HN points • 03 Nov 24
🇺🇸 U.S. Politics Forecasting
  1. A professor predicts that Donald Trump has a greater than 60% chance of winning the 2024 U.S. presidential election. This prediction is based on computer simulations rather than traditional polling.
  2. The simulations suggest Trump will likely win key states like Michigan, Ohio, and Florida, while Harris is expected to win states like Georgia and Arizona.
  3. The forecasting method used is known as Agent-Based Modeling, which combines real data about voters and economic conditions to make predictions rather than relying on expert opinions.

What people are not seeing about TimeGPT

Three Data Point Thursday • 19 implied HN points • 16 Nov 23
🕹 Technology Forecasting
  1. Time series models, like TimeGPT, are advancing and will provide a significant boost in machine learning capabilities.
  2. Adding time as a feature in models can enhance data analysis due to the information richness of recent data.
  3. Although skepticism exists around time series machine learning models, advancements in generic models like TimeGPT are removing some barriers.

Preliminary 2025 Housing Forecasts

CalculatedRisk Newsletter • 38 implied HN points • 22 Nov 24
💰 Finance Forecasting
  1. Housing prices are expected to rise a little in 2025, but not by much. People think the increase will be in the low to mid single digits.
  2. Sales of new and existing homes are predicted to go up next year. However, existing home sales will likely stay around four million.
  3. The construction of multi-family homes is not expected to improve in 2025. Builders are generally cautious about starting new multi-family projects.

Don't Run Coibion-Gorodnichenko Regressions with Micro Data

inexactscience • 39 implied HN points • 27 Mar 23
💰 Finance Forecasting
  1. Running Coibion-Gorodnichenko regressions with individual data can lead to misleading results. It's important to use appropriate data types to avoid confusion in the findings.
  2. Individual forecasts tend to produce negative results compared to positive results in average forecasts. This means that the insights from these regressions can differ significantly based on the data used.
  3. The methodology is sensitive to noise and measurement errors, which can skew results. Researchers need to be cautious and robust in their approach to ensure accurate interpretations.

When to Use Forecasting for Decisions

Abstraction • 19 implied HN points • 13 Dec 24
💼 Business Forecasting
  1. It's not always worth it to forecast when making decisions. Sometimes it's better to prepare for the worst or trust experts who know what they're doing.
  2. For less important choices, you can follow proven rules or experts. This makes decision-making easier and saves time.
  3. When facing big decisions, like moving cities, it's smart to gather data to guide your choice. Using information about others’ experiences can help you make better decisions.

Forecasting, Game QA, and Personalized Embodied Agents

ppdispatch • 8 implied HN points • 30 May 25
🕹 Technology Forecasting
  1. A new type of learning called outcome-based reinforcement learning is helping smaller language models make accurate predictions, even better than some big models.
  2. Researchers are looking at how AI agents remember information to provide personalized help, but they still struggle with remembering complex user preferences.
  3. A new benchmark for video game testing helps measure how well AI models can find bugs and glitches in games, making the testing process better and more efficient.

Data matters

Klement on Investing • 3 implied HN points • 13 Aug 25
💰 Finance Forecasting
  1. Using high-quality data sources can really improve an analyst's accuracy. It’s better to have various data subscriptions than just relying on experience.
  2. Smaller, specialized data providers have a bigger positive impact on forecast accuracy than larger firms. But these smaller providers often get cut for cost-saving reasons.
  3. Cutting back on data access can harm investment research. It can lead to less unique insights, making it hard for analysts to stand out, which isn't good for business.

Wanna run 50,000 election sims in 2 seconds?

Matt’s Five Points • 19 implied HN points • 04 Nov 22
🇺🇸 U.S. Politics Forecasting
  1. You can run a quick election simulation by using an Excel sheet. Just change the win probabilities for each state and the sim does the math for you in about 2 seconds.
  2. Basic election modeling isn't as hard as it sounds. You can easily create your own model with some data and a few calculations to forecast election outcomes.
  3. Strong, accurate models take more work and understanding, but anyone can start trying their hand at it. It can be enjoyable to explore different scenarios with the data.

Let me fix those forecasts for you

Klement on Investing • 7 implied HN points • 03 Jan 25
💰 Finance Forecasting
  1. Forecasts for stock market returns are often inaccurate. For example, analysts expected the S&P 500 to rise by about 8% in 2024, but it actually rose by 23%.
  2. Historical data shows a low correlation between predicted and actual stock market returns. Over the last 20 years, the correlation for analysts' forecasts has been very weak.
  3. Using forecast errors, we can adjust predictions for the next year. For 2025, the S&P 500's return could realistically range from a 29% drop to a 47% increase.

Identifying Top Forecasters

Abstraction • 24 implied HN points • 17 Sep 23
🔬 Science Forecasting
  1. Identifying top forecasters is valuable for preparing for the future
  2. Distinguishing between luck and skill in forecasting is a challenge
  3. Factors like breadth vs depth and domain bias play a role in ranking top forecasters

Leviathan wakes

Gradient Ascendant • 16 implied HN points • 21 Feb 24
🕹 Technology Forecasting
  1. The author quit their job to work on a new AI-related project motivated by the transformative potential of modern AI technology.
  2. Google's Gemini 1.5 model is a significant advancement in AI capabilities, able to handle an impressive 10 million tokens for input, marking a major leap forward in AI development.
  3. Despite its imperfections, Gemini 1.5 and other advanced AI models are drastically reducing limitations and opening up new possibilities for future technological innovations.

Chuyện Việt Nam: Forbes vinh danh người dự báo "đến đầu tháng 8/2021 Sài Gòn sẽ hết dịch"

Thái | Hacker | Kỹ sư tin tặc • 19 implied HN points • 21 Feb 22
🏥 Health Politics Forecasting
  1. Ngô Hoàng Anh and team accurately predicted the end of the COVID-19 outbreak in Saigon by August 2021 using their SEIQHCDRO model.
  2. Collaboration with the Epidemiological Modelling Unit ensured adjustments to their model for accurate COVID-19 predictions in Saigon.
  3. Future forecasts by the team suggest a potential new wave of COVID-19 in Saigon from December 2021 to March 2022, depending on the enforcement of preventive measures.

Forecast Scoring Methods

Abstraction • 19 implied HN points • 07 Nov 23
🔬 Science Forecasting
  1. Market-Based Platforms are good for gauging market sentiment, not individual forecasting skill
  2. Reputation-Based Platforms focus on individual performance metrics to identify top forecasters
  3. Consider the ramifications of overconfidence when selecting a scoring system for forecasting

Scoring Rules

Abstraction • 19 implied HN points • 26 Sep 23
🔬 Science Forecasting
  1. Proper scoring rules encourage honest and accurate forecasting by penalizing dishonesty and over/under-confidence.
  2. Improper scoring rules do not incentivize forecasters to report their true beliefs, leading to suboptimal forecasting incentives.
  3. In practice, proper scoring mechanisms like Brier scoring help distinguish skill from noise over multiple rounds and promote honest, calibrated forecasting.

Silver Bulletin pollster ratings 2024 archive

Silver Bulletin • 4 implied HN points • 17 Feb 25
🇺🇸 U.S. Politics Forecasting
  1. The Silver Bulletin has updated its pollster ratings for the 2024 presidential election, also keeping the old ones for reference. This helps everyone see how reliable different pollsters are.
  2. Not many new polls have come in since the last update, so the overall ratings haven't changed much. The most accurate pollsters from previous years are still at the top.
  3. The 2024 polling data includes some recent elections, but the accuracy of the polls has been similar to past years, with a noticeable bias in some cases. It's important to stay aware of this while following the elections.

The Insufficient Weirdness Hypothesis

Gradient Ascendant • 20 implied HN points • 01 Jun 23
🔬 Science Forecasting
  1. The future is consistently weirder than expected because of unknown unknowns and unusual juxtapositions.
  2. AI development and outcomes are expected to be highly weird and unpredictable, not following a smooth exponential path.
  3. Weird and unexpected scenarios are more indicative of potential future risks to consider rather than conventional outcomes.

Predicting the Future

The Cosmopolitan Globalist • 4 implied HN points • 02 Jan 25
🇺🇸 U.S. Politics Forecasting
  1. Predicting the future is hard because people often think they're better at it than they really are. Many make mistakes like being too confident or not paying attention to the facts.
  2. Some people, called superforecasters, are really good at making accurate predictions. They focus on facts, break problems into smaller parts, and keep track of their past predictions to learn from them.
  3. To improve predictions, it's important to express guesses as probabilities, look for different opinions, and stay open to changing your mind when new information comes in.

Gradient Flow #39: Becoming TikTok, Next-gen Workflow Orchestration and Forecasting

Gradient Flow • 19 implied HN points • 15 Jul 21
🕹 Technology Forecasting
  1. The newsletter discusses next-gen dataflow orchestration and automation systems like Prefect, a startup that helps manage dataflows.
  2. It introduces cool new open source tools like Greykite, a flexible and fast library for time-series forecasting.
  3. BytePlus, a new division of ByteDance, is offering the technology behind TikTok to websites and apps, presenting interesting challenges in the global market.

Results for the Salem/CSPI Prediction Tournament

Center for the Study of Partisanship and Ideology • 11 implied HN points • 29 Aug 23
💰 Finance Forecasting
  1. The winners of the Salem/CSPI Prediction Tournament were announced, including a $25,000 prize and a fellowship
  2. The analysis of the betting markets showed mixed results, with some events being accurately predicted while others were not
  3. Participants in the tournament were mostly young, male, and had a libertarian-leaning political orientation

Do you trust AI forecasts?

Klement on Investing • 1 implied HN point • 09 Jul 25
💰 Finance Forecasting
  1. Many people trust human predictions over AI forecasts, but this can change based on a person's knowledge of AI. Those who understand AI better often trust its predictions more than those from humans.
  2. Some investors believe AI forecasts can be more reliable than human forecasts because they see flaws in human analysis. This creates a divide between traditional and tech-savvy investors.
  3. There are different styles of analysts: some focus on numbers and models while others rely on qualitative insights. Different investors prefer different styles based on their own backgrounds and preferences.

Oh those share price forecasts

Klement on Investing • 6 implied HN points • 03 Jan 24
💰 Finance Forecasting
  1. Analyst forecasts and strategist forecasts often differ in predicting market outcomes.
  2. Strategists tend to be more pessimistic while analysts are generally overly optimistic.
  3. Relying on market forecasts for investment decisions may not be the best approach as accuracy is often low.

Stars are treated differently

Klement on Investing • 1 implied HN point • 17 Dec 24
💰 Finance Forecasting
  1. Star analysts have a bigger influence than regular analysts because their reports can move markets. They are rewarded for giving high-quality forecasts that help investors succeed.
  2. To become a star, analysts need to make good predictions, but staying a star often means making bold and risky calls instead of focusing on quality.
  3. Once an analyst becomes a star, they are judged less harshly for mistakes. This means they can take more risks and make big headlines, even if it doesn't always lead to good outcomes.

AI Towns and Agent Based Modeling

Magis • 3 HN points • 26 Aug 23
🕹 Technology Forecasting
  1. Agent-based modeling uses computer agents to simulate interactions and behavior based on rules.
  2. Large Language Models (LLMs) could enhance agent-based modeling by providing agents with realistic context and knowledge.
  3. Improved agent-based modeling could revolutionize economic forecasting by simulating population-level effects and simplifying forecasting.

New $5,000 Prize in the Salem Center/CSPI Forecasting Tournament

Center for the Study of Partisanship and Ideology • 3 implied HN points • 15 Feb 23
🇺🇸 U.S. Politics Forecasting
  1. A new $5,000 prize is added to the Salem Center/CSPI Forecasting Tournament to encourage more participation.
  2. The prize will go to the person with the greatest percentage increase in their portfolio value by the end of the tournament.
  3. Various markets on global events and economic forecasts are also discussed in the tournament.