The hottest Forecasting Substack posts right now

And their main takeaways
Category
Top Business Topics

Scribble-based forecasting and AI 2027

DYNOMIGHT INTERNET NEWSLETTER • 562 implied HN points • 30 Jun 25
đź•ą Technology Forecasting
  1. Both math and intuition can be used for forecasting, but they serve different purposes. Sometimes, using intuition can be more practical when creating predictions about complex situations.
  2. Math-based forecasts are best when the rules of a situation are well understood and complex. For simpler scenarios, basic predictions may be just as effective.
  3. Creating simple visual predictions, like drawing lines, can help clarify your thoughts. It's a great exercise to explore different potential outcomes and express predictions clearly.

✨ The Age of AI: Will 2025 be our last normal year?

Faster, Please! • 1279 implied HN points • 03 Jan 25
đź•ą Technology Forecasting
  1. AI technology is rapidly evolving, and some predict it could change our everyday lives significantly by 2025. If this happens, what we consider 'normal' now might no longer exist.
  2. Recent advances in AI, like OpenAI's new model, have made experts rethink how soon we might see 'strong' AI that can perform complex tasks like humans. This raises important questions about the future of work and society.
  3. Despite the excitement around AI, not all experts believe we are close to seeing a major economic boom from it. Predictions about technology can be tricky, and history shows change can take a long time.

The U.S. Economy is Roaring. Why Should Rates Come Down?

The Overshoot • 452 implied HN points • 27 Jan 24
đź’° Finance Forecasting
  1. The U.S. Economy is showing strong growth and may not need rate cuts despite controlled inflation.
  2. Traders anticipate interest rates to decrease, but data suggests a period of faster growth akin to past economic booms.
  3. Initial forecasts of a U.S. recession were proven wrong, with the economy growing over 3% and showing resilience against negative predictions.

Chinese Hydropower - Large and Underappreciated

Syncretica • 648 implied HN points • 06 Nov 23
đź’° Finance Forecasting
  1. China has the largest hydropower sector globally, with a significant impact on power generation worldwide.
  2. Hydropower output is heavily influenced by weather conditions, with recent rainfall improvements expected to boost Chinese hydropower production.
  3. The strong growth in Chinese hydropower output is likely to lead to a decrease in fossil fuel imports and a reduction in thermal power generation.

My Feast-or-Famine AI Bet with Holden Karnofsky

Bet On It • 296 implied HN points • 21 Jul 25
đź•ą Technology Forecasting
  1. Holden believes AI will greatly change the economy, but he isn't sure if it will be for the better or worse. Bryan thinks that we won't see these big changes for a long time, maybe decades.
  2. They made a bet about the future economy, betting on whether AI will boost or damage the global economy by 2044. If the economy is either much better or much worse than it is now, Holden wins; otherwise, Bryan wins.
  3. Bryan will decide the winner of the bet, but they agreed on backup judges in case he can't. This shows there's trust between them in this friendly wager.
Get a weekly roundup of the best Substack posts, by hacker news affinity:

Art to Science

Abstraction • 39 implied HN points • 02 Jan 26
đź•ą Technology Forecasting
  1. Forecasting bots can run continuously, answer many questions, and be scored in real time, turning forecasting from a slow craft into a fast, repeatable process.
  2. Large, scored tournaments and shared datasets will let people empirically test different methods and finally learn which forecasting approaches actually work at scale.
  3. Simple heuristics get you most of the way there, but reaching the frontier requires deeper techniques and open sharing of methods to accelerate progress.

Pre-Mortems and Sanity Checks

Abstraction • 29 implied HN points • 14 Jan 26
đź•ą Technology Forecasting
  1. Do a pre-mortem: assume the forecast is wrong and list plausible ways it could fail (like cancellations, acquisitions, or shifted definitions) so you don’t miss important paths.
  2. Run a sanity check to make sure the probability fits basic world knowledge and common sense, and correct obvious errors like using the wrong base rate.
  3. Make these checks the final gate: if either one flags a problem, rework the forecast or use a different approach before submitting.

The Broken Leg Check

Abstraction • 34 implied HN points • 07 Jan 26
đź•ą Technology Forecasting
  1. Do a quick "broken leg" check first because a decisive news event can resolve a question immediately and save the time and cost of running the full forecasting pipeline.
  2. Be cautious: a wrongly triggered broken-leg update is dangerous since proper scoring heavily penalizes confident incorrect forecasts, so false positives can wipe out gains.
  3. Treat it as an empirical trade-off: implement a news-based detector, clearly define what "overwhelmingly resolves" means, track when it fires, and tune thresholds, confidence damping, or disable it if blowouts outweigh the savings.

Reconsidering the "2020s as 1940s" Analogy

The Overshoot • 550 implied HN points • 08 Mar 23
đź’° Finance Forecasting
  1. The global economy faced crises in different time periods, revealing government responses can impact recovery.
  2. Excessive debts before a crisis can hinder growth post-crisis, affecting employment and national income.
  3. Governments borrowing and spending during emergencies can lead to positive outcomes, improving sectors and reducing debt burdens.

On the Political Economy of Mendacity in 21st Century America

Brad DeLong's Grasping Reality • 330 implied HN points • 10 Jun 25
🇺🇸 U.S. Politics Forecasting
  1. Economic analysis can sometimes feel like a performance rather than a serious discussion. People in media might act like they believe things that aren't true.
  2. Immigration can boost a country's economy, but some policies can slow down growth. If certain negative policies continue, economic growth could be very low.
  3. Even when people know the truth about economic issues, they still might not say it. This creates a situation where misleading information can seem credible.

A random number generator determined the “favorite" in our forecast

Silver Bulletin • 905 implied HN points • 05 Nov 24
🇺🇸 U.S. Politics Forecasting
  1. The presidential race is extremely close, almost a 50/50 chance for each candidate. This makes it tough to predict who will win.
  2. Recent simulations showed a slight edge for Kamala Harris, but overall results were very mixed, indicating a highly uncertain outcome.
  3. Voting is crucial in this tight race; it really comes down to what people choose, making it more important than any guessing games.

Decay Functions

Abstraction • 29 implied HN points • 09 Jan 26
đź•ą Technology Forecasting
  1. A single probability for a time window needs a decay model because where the probability mass sits across the window determines how much chance remains as time passes.
  2. Probability can follow different hazard patterns—constant (linear decay), increasing (back-loaded, like last‑minute negotiations), decreasing (front‑loaded, like ceasefires), or event‑driven—and each pattern changes how fast the cumulative probability is consumed over time.
  3. The forecasting bot classifies which hazard applies (defaulting to constant when unsure) and uses that to update remaining probability as time elapses, but this is a refinement that can be misclassified and matters most for long‑horizon questions.

Classification and Method Selection

Abstraction • 29 implied HN points • 08 Jan 26
đź•ą Technology Forecasting
  1. Match the forecasting method to the question type: classify questions into base-rate, time-series, conditional-chain, or novel-event and route each to a specialized approach.
  2. Use the right technique for each class: use historical reference classes and adjustments for base rates, simulate trajectories for time-series questions, multiply conditional probabilities for conjunctive chains, and apply a Laplace-style prior for unprecedented events.
  3. Track and improve empirically: use an LLM classifier (defaulting to base rate when unsure), choose reference classes and decompositions carefully, and measure which methods are over- or under-confident as you scale.

What will 2026 bring?

Erdmann Housing Tracker • 21 implied HN points • 21 Jan 26
đź’° Finance Forecasting
  1. Metro-area analyses act as 'all else equal' forecasts, so they project outcomes assuming other factors don’t change while still needing to account for many variables.
  2. A near-zero 2025 home-price forecast (about 0.1%) matched the observed change in the Zillow Home Value Index, showing that small, precise forecasts can be accurate.
  3. The outlook for 2026 calls for roughly 3% home-price appreciation, even though expert forecasts for 2025 varied widely from about -2% to 10.8%.

Deploying a Forecasting Bot

Abstraction • 29 implied HN points • 05 Jan 26
đź•ą Technology Forecasting
  1. A structured, reproducible forecasting pipeline models how strong human forecasters think so methods can be tested and refined systematically.
  2. Huge cost cuts made iteration affordable: per-question cost dropped from $0.109 to $0.004 (about 27Ă—), enabling many more experiments across the tournament.
  3. The team accepts a likely short-term performance hit by using cheaper models and fewer tokens because the priority is learning which pipeline parts truly matter using the tournament as a feedback loop.

The great AI forecasting divide

Faster, Please! • 548 implied HN points • 15 Feb 25
đź•ą Technology Forecasting
  1. There is a debate about whether AI will change society in a big way or just a small one. Some experts think it could be revolutionary, while others see it as an evolution of technology.
  2. Economists base their predictions about AI on how past technologies have changed society. They might not expect the rapid advances that could happen sooner than anticipated.
  3. The discussion about AI's impact raises questions about our future and how quickly we might see changes in our lives and jobs because of intelligent machines.

The polls are close, but that doesn’t mean the results will be

Silver Bulletin • 666 implied HN points • 04 Nov 24
🇺🇸 U.S. Politics Forecasting
  1. Polls show a very tight race, but this doesn't guarantee a close election outcome. Both candidates could end up winning by a larger margin than the polls suggest.
  2. Polling errors can happen in either direction, making the election unpredictable. Even a small error could lead to a surprising result on election day.
  3. The focus should be on battleground states, as those will ultimately determine the winner regardless of national polls. The voting patterns in these key areas could lead to differing results from the national poll averages.

Building Models to Predict the Future

Growth Croissant • 353 implied HN points • 31 Mar 23
đź’Ľ Business Forecasting
  1. Subscription business models are more predictable and easier to forecast compared to other models.
  2. A reliable forecast can provide baseline expectations for growth, shift focus to long-term revenue, and help set goals and allocate investments.
  3. Forecasting new paid subscriptions and predicting paid cancels are key components in building a forecast model for a subscription business.

Talkin’ bout an AI Revolution: 3 scenarios

Faster, Please! • 365 implied HN points • 26 Feb 25
đź•ą Technology Forecasting
  1. By 2030, we might still be at the start of a major AI development period. It's okay because this means we have a lot of exciting advancements ahead.
  2. More traditional institutions, like big banks, are now seriously talking about AI. This shows that AI is becoming a big deal in the mainstream world, not just in tech circles.
  3. Experts believe that as AI keeps getting better, the 2020s could see various new economic and technological changes. This could change how we live and work in many ways.

SBSQ #14: Election Day lightning round!

Silver Bulletin • 507 implied HN points • 05 Nov 24
🇺🇸 U.S. Politics Forecasting
  1. Election Day tends to be calm since there’s not much to do until results come in. It's a good time to reflect instead of focusing too much on exit polls.
  2. Different prediction models, like FiveThirtyEight's, can have varying odds for candidates which might not reflect the true situation. It's important to pay attention to both polls and the underlying fundamentals.
  3. There are concerns that too many prediction models can lead pollsters to stick closely to common predictions, impacting the variety of polling results we see.

Why it doesn't pay to make predictions

The Cosmopolitan Globalist • 16 implied HN points • 03 Jan 26
🌍 World Politics Forecasting
  1. Predictions often fail, so it's wiser not to make firm forecasts when information is limited.
  2. There is a genuine hope that Venezuela can be stabilized quickly and become democratic and prosperous, but the outcome is uncertain.
  3. Global events can be wildly surprising, and while dramatic scenarios are tempting to imagine, it's better to admit uncertainty than pretend to know the future.

Moneyballocracy

Gradient Ascendant • 20 implied HN points • 22 Dec 25
đź•ą Technology Forecasting
  1. AI models are rapidly getting good at forecasting and already rival the wisdom of crowds and some human forecasters.
  2. Forecasting with AI is cheap and scalable, so you can run detailed, conditional predictions across thousands of stocks, counties, or scenarios that used to be impractical.
  3. Making the future more legible will reshape elections and politics: it can help match policy to voter preferences but also enable targeted manipulation, and any side that uses it effectively will gain a real advantage.

Artificial Intelligence Weather Model AIFS

Open-Meteo • 843 implied HN points • 29 Feb 24
đź•ą Technology Forecasting
  1. ECMWF released its cutting-edge artificial intelligence weather model AIFS as open-data, marking a significant move in the open-data weather forecasting landscape.
  2. AIFS uses Graph Neural Networks to learn complex weather patterns, showcasing superior accuracy in longer-range forecasts exceeding 5 days.
  3. While AIFS has limitations in weather variables range and interval forecasts, its open availability enables users to compare its forecasts with traditional models, offering a new perspective in weather forecasting.

AI and the future of weather forecasting

The PhilaVerse • 123 implied HN points • 02 Jul 25
đź•ą Technology Forecasting
  1. AI is changing how we predict the weather by offering quicker and more efficient methods compared to traditional forecasting. This helps provide better updates, especially for things like storms and heatwaves.
  2. While AI forecasting models are fast, they currently work at a lower resolution than traditional systems. They still depend on traditional methods for some accurate initial data.
  3. There is growing interest worldwide in using AI for weather forecasting. This technology could improve disaster preparedness, agriculture, and energy management, making it valuable for many industries.

Question #8 for 2026: How much will Residential investment change in 2026? How about housing starts and new home sales in 2026?

CalculatedRisk Newsletter • 14 implied HN points • 26 Dec 25
đź’Ľ Business Forecasting
  1. Residential investment is likely to be down year‑over‑year in 2026.
  2. Total housing starts are expected to fall slightly as multi‑family starts decline while single‑family starts remain mostly unchanged.
  3. New home sales are likely to be largely unchanged year‑over‑year, though forecasts are uncertain because three months of data are missing and units under construction remain above pre‑pandemic levels.

Older people consume differently

Klement on Investing • 4 implied HN points • 03 Feb 26
đź’Ľ Business Forecasting
  1. People buy different things as they age: healthcare and housing spending tends to rise while education, leisure, clothing and transport fall.
  2. The pattern depends on national demographics. Similar population declines can lead to very different sector effects — Japan shows broad declines, China has healthcare holding up while leisure and transport fall sharply, and Singapore mixes increases and declines because it is ageing but still growing.
  3. This shifts the revenue outlook for companies: leisure and clothing retailers face structural declines while healthcare providers and food retailers look more resilient.

The Pipeline Metrics That Matter

Bottom Up by David Sacks • 281 implied HN points • 29 Oct 24
đź’Ľ Business Forecasting
  1. Tracking pipeline generation is crucial for growth in SaaS companies. If new opportunities are increasing, it's a good sign to hire more sales staff; if not, boost marketing efforts.
  2. Understanding pipeline conversion metrics helps identify where improvements are needed. Knowing how long deals take to close and where they tend to get stuck can lead to better sales processes.
  3. Active pipeline metrics allow for accurate forecasting. Keeping an eye on open opportunities and their expected close dates helps businesses plan and strategize effectively.

A Different Approach to Revenue Estimates Leveraging Alternative Data

The Data Score • 138 implied HN points • 24 May 23
đź’° Finance Forecasting
  1. Leveraging alternative data for revenue estimates goes beyond traditional transaction data, focusing on customer acquisition and retention insights.
  2. Applying the customer acquisition funnel framework to alternative data can help identify early trends and potential growth issues in a business.
  3. Monitoring the journey from awareness to loyalty using alternative data sets can offer valuable insights for predicting sustainable revenue growth beyond the short term.

Silver Bulletin pollster ratings 2025 archive

Silver Bulletin • 6 implied HN points • 14 Jan 26
🇺🇸 U.S. Politics Forecasting
  1. Pollsters are ranked by historical accuracy and transparency using a Predictive Plus-Minus score that is converted to letter grades. A negative plus-minus means the pollster is expected to be more accurate than average.
  2. The ratings use multiple measures — simple and advanced plus-minus, mean-reverted bias, house effects, and an ADPA herding penalty — and give bonuses for transparency like AAPOR or Roper Center sharing. These metrics together adjust for sample size, timing, and how a poll compares to others.
  3. The archive was updated with hundreds of new polls from the 2024 presidential, congressional, and gubernatorial elections, and full datasets (pollster stats and raw polls) are available for download. The update shifted some ratings but the top pollsters remained largely the same.

Advancing Beyond Short-Term Solar Forecasts [Pt.2]

Space Ambition • 79 implied HN points • 08 Dec 23
🔬 Science Forecasting
  1. It's important to understand the solar cycle better and predict solar storms. These storms can cause big financial losses and affect many technologies we rely on.
  2. Currently, we can only accurately predict space weather for about three days ahead. This is because solar events happen quickly, and predicting them is really complicated.
  3. We need more advanced tools and methods, like machine learning, to improve our predictions. Using new technology can help us learn more about the Sun and its effects on Earth.

March Madness Forecasts: FiveThirtyEight vs. Betting Markets

Mike’s Blog • 78 implied HN points • 07 Apr 23
đź’° Finance Forecasting
  1. Betting markets slightly outperformed FiveThirtyEight in predicting NBA, NFL, and MLB games.
  2. New data collected for March Madness shows both FiveThirtyEight and betting markets performed similarly, and neither significantly outperformed.
  3. Hypothesis: Both betting markets and experts may have worse accuracy in playoffs and tournaments compared to regular season games.