The hottest Forecasting Substack posts right now

And their main takeaways
Category
Top Business Topics

SBSQ #30: Will liberals turn against sports betting?

Silver Bulletin • 293 implied HN points • 23 Mar 26
🇺🇸 U.S. Politics Forecasting
  1. He doesn’t incorporate prediction market prices into his models because he wants to preserve an independent perspective and avoid pre-diluting or effectively betting against his own forecasts.
  2. Prediction market inputs create technical problems — they can cause recursion and feedback loops and are highly correlated with existing signals, which makes models unstable and hard to fit robustly.
  3. Sports betting is drawing bipartisan backlash, with lawmakers and public figures pushing restrictions and many people annoyed by the ads and industry influence, signaling potential political momentum for regulation.

What Happened With Bio Anchors?

Astral Codex Ten • 12044 implied HN points • 12 Feb 26
🕹 Technology Forecasting
  1. A compute-centered forecasting approach correctly captured that AI progress has largely tracked available compute and scaling laws, which explains much of the recent boom.
  2. The main error was underestimating algorithmic progress and effective compute growth (including longer training runs and test-time compute), so systems became far more powerful each year than the model assumed and pushed timelines much earlier.
  3. Forecasts are still useful but hinge on a few sensitive parameters, so you need proper sensitivity analysis and humility — uncertainty can cut both ways and make outcomes riskier than naive skepticism assumes.

How Accurate Are Learning Curves?

Construction Physics • 19208 implied HN points • 24 Dec 25
🕹 Technology Forecasting
  1. Learning rates often change over time and many cost-versus-production curves show breakpoints instead of a single straight line on a log–log plot.
  2. Early learning rates are weak predictors of later learning rates, so using a single historical rate to forecast future costs is unreliable.
  3. Allowing learning rates to change probabilistically (piecewise models) can improve forecasts for some technologies, but the gains are modest and depend on the product, so combining probabilistic outside-view methods with technology‑specific inside‑view analysis is most useful.

COOPER NCAA basketball ratings

Silver Bulletin • 149 implied HN points • 18 Mar 26
🎾 Sports Forecasting
  1. COOPER is a new power-rating system that ranks all 365 Division I men’s basketball teams using wins and losses, margin of victory, team tempo, preseason polls, and conference strength.
  2. The ratings are Bayesian/Elo-like and update continuously: an impact factor weights recent games, close matchups, conference games, and NCAA tournament games more heavily, and ratings partly carry over between seasons.
  3. COOPER offers detailed outputs (offensive/defensive ratings, strength of schedule, home-court factors), an objective-only variant, historical season-ending ratings back to 1950, and tools to convert ratings to win probabilities; tournament forecasts blend COOPER with other models and injury data and are available to subscribers.

Mantic Monday: The Monkey's Paw Curls

Astral Codex Ten • 14109 implied HN points • 13 Jan 26
🔮 Crypto Forecasting
  1. Prediction markets have exploded in volume and produce accurate probabilities, but most activity is degenerate gambling and they haven’t yet changed how society or the media make decisions.
  2. Vague resolution rules and decentralized oracles cause frequent disputes, insider trading concerns, and "rulescuck" losses, and proposed technical fixes (like using LLMs) carry their own risks.
  3. Conditional "decision" markets could be transformative if they can avoid confounding — one proposed fix is markets that predict the eve-of-decision market prices — and AI superforecasters may soon supplant human markets, leaving either better user-driven platforms or AI-led forecasting as the likely path forward.
Get a weekly roundup of the best Substack posts, by hacker news affinity:

BREAKING: The AI 2027 doomsday scenario has officially been postponed

Marcus on AI • 14742 implied HN points • 21 Nov 25
🕹 Technology Forecasting
  1. The high-profile "AI 2027" doomsday prediction has been postponed, and AGI is unlikely to arrive in 2027 and probably not this decade.
  2. National policy and big parts of the economy were built around the assumption of imminent AGI, so those plans and investments need to be seriously rethought.
  3. The doomsday narrative was largely speculative and served as marketing, amplified by media and influencers while dissenting views were downplayed, showing we relied too much on hype instead of sober analysis.

COOPER women's NCAA basketball ratings

Silver Bulletin • 86 implied HN points • 15 Mar 26
🎾 Sports Forecasting
  1. Final regular-season COOPER power ratings have been published for all 363 Division I women's basketball teams, tracking each team’s highs and lows over the year.
  2. COOPER is a new Elo/Bayesian-style rating system that blends wins, margin of victory, tempo, preseason polls, and conference strength, and it weights recent, close, conference, and tournament games more heavily with some customization for the women’s game.
  3. Paid subscribers get the full dataset and extras — offensive/defensive ratings, strength of schedule and home-court factors, an objective-only version, historical season-end ratings back to 2002–03, a spreadsheet to convert ratings into win probabilities, and NCAA tournament forecasts coming after the brackets.

ACX/Metaculus Prediction Contest 2026

Astral Codex Ten • 4060 implied HN points • 27 Dec 25
📰 News Forecasting
  1. A crowdsourced prediction contest on Metaculus is now live, covering U.S. politics, AI, international affairs, and culture, and you can enter using your regular account or a bot account.
  2. Submit forecasts by January 17 at 11:59 PM PT; a snapshot then determines contest rankings and how the $10,000 prize pool is allocated, and forecasts made after that only affect site leaderboards, not contest rankings.
  3. Organizers announced cash awards for the best question submitters, with the top prize being $700 and several other winners receiving smaller amounts.

Which 2028 Democrats have the best electoral track record?

Silver Bulletin • 473 implied HN points • 20 Feb 26
🇺🇸 U.S. Politics Forecasting
  1. Gavin Newsom has never lost statewide in California, but his victory margins are consistently smaller than the margins Democrats get in presidential years, so he underperforms the presidential benchmark.
  2. The Simple Benchmark (SB) Score compares a Democrat’s margin to the Democratic presidential baseline in their state or district, and it adjusts for non-presidential years using nearby presidential results and shifts in the House popular vote.
  3. Several Democrats — notably Andy Beshear, Amy Klobuchar, Ruben Gallego, Gretchen Whitmer and Josh Shapiro — post much stronger SB Scores, indicating they have better electoral track records and may be more electable than Newsom.

Will the Housing Market Crash in 2026?

Progress and Poverty • 1885 implied HN points • 13 Jan 26
💰 Finance Forecasting
  1. An 18-year land-cycle theory says fixed land supply makes real estate unusually prone to recurring speculative booms and busts driven by credit, building cycles, and expectations about future resale values.
  2. The historical pattern is suggestive but weak: the data set is small, several peaks require retrofitting to fit the 18‑year story, and market timing is generally unreliable, so the model is not a strong tool for precise investment forecasts.
  3. Recent housing indicators—high price-to-rent, a large real-estate share of GDP, falling affordability, and elevated new-home inventory—match the theory’s warning signs but differences from 2008 mean a crash is uncertain; the theory nonetheless implies that land-value taxation could dampen speculation and crises often create windows for policy reform.

Strategy in four worlds

Subconscious • 1265 implied HN points • 11 Jan 26
🕹 Technology Forecasting
  1. Risk and uncertainty are different: risk is measurable and fits expected-utility tools, while uncertainty involves unknown possible outcomes and needs a different approach. You can categorize environments as clear, complicated, complex, or chaotic based on how cause and effect behave.
  2. Match your tactics to the environment: clear and complicated problems reward forecasting, expert analysis, and optimization, whereas complex systems require robust, antifragile strategies that map feedback loops, and chaotic situations demand fast reflexes and simple orientation to survive.
  3. Scenario planning is the right tool for complexity: it helps identify major drivers, surface feedback loops, and wind‑tunnel strategies across many plausible futures so you can build robustness or intentionally shape outcomes. Because real challenges mix these worlds, skilled strategists combine forecasting, scenarios, and adaptive judgment rather than relying on one model.

Can a Transformer “Learn” Economic Relationships?

Arpitrage • 2194 implied HN points • 22 Dec 25
🕹 Technology Forecasting
  1. Transformer-based models can learn the dynamics of a New Keynesian economy from simulated data and produce accurate out-of-sample forecasts, outperforming simple reduced-form benchmarks.
  2. They often predict the direction and rough magnitude of policy shock responses, but misestimate impulse-response dynamics and can exhibit overshooting, so they do not fully recover the true causal structure.
  3. These advances weaken the practical bite of the Lucas critique by improving prediction, but they do not eliminate the need for structural models for causal interpretation, welfare analysis, and interpretability; transformer methods are a promising complementary tool.

Suggest Questions For Metaculus/ACX Forecasting Contest

Astral Codex Ten • 3441 implied HN points • 14 Nov 25
🇺🇸 U.S. Politics Forecasting
  1. The ACX and Metaculus forecasting contest is looking for interesting questions related to events in 2026. Questions should have clear answers that can be known by the end of the year.
  2. The top ten contributors of question ideas can win prizes ranging from $150 to $700. It's a chance to get recognized for your ideas and earn some money.
  3. This year, the contest will also include AI bots that will compete against human forecasters. This adds a fun twist and a different level of competition to the event.

The Paris Delusion

The Honest Broker Newsletter • 2070 implied HN points • 22 Nov 25
🌞 Climate & Environment Forecasting
  1. Saying the Paris Agreement alone caused a big drop in projected warming is misleading; the apparent improvement mostly reflects earlier scenarios that over‑predicted coal use and were therefore wrong, not clear policy-driven emissions cuts.
  2. Actual data show no acceleration in global decarbonization since Paris: emissions per unit of GDP have fallen at about 2% per year, far below the roughly 8% per year sustained cuts needed for deep decarbonization and never achieved by any country.
  3. We need honest, evidence‑based policymaking — stronger and effective measures to speed real decarbonization are required, while also protecting energy access, supply reliability, and affordability, instead of celebrating questionable success stories.

‘Russians hungry but not starving’

Wrong Side of History • 460 implied HN points • 21 Jan 26
📰 News Forecasting
  1. Journalists and media pundits often make attention-grabbing predictions and are frequently wrong because they have no skin in the game and have strong ideological biases.
  2. Predicting foreign policy is especially hard since it depends on culture, personalities, and many interacting factors, so disciplined non-specialists (superforecasters) can sometimes outperform supposed experts.
  3. Even respected newspapers and intellectuals can badly misjudge major events — for example, influential commentators once praised Ayatollah Khomeini and underestimated the dangers of the Iranian revolution.

Who’s ahead on the generic congressional ballot?

Silver Bulletin • 379 implied HN points • 04 Feb 26
🇺🇸 U.S. Politics Forecasting
  1. Democrats hold a modest lead of about D +5.5 on the generic congressional ballot, up from roughly D +3 between June and November.
  2. Individual polls vary a lot — results this week ranged from about D +1 to D +9 — but the average smooths those swings and weights polls by pollster quality, sample size, recency, and frequency while preferring likely-voter samples.
  3. Many of the polls in the average were conducted before the Jan. 24 killing of Alex Pretti, so subsequent public reaction could push the generic ballot further toward Democrats, and paid subscribers can access state benchmarks and historical generic-ballot averages back to 1994.

Introducing COOPER: Silver Bulletin's NCAA basketball rating system

Silver Bulletin • 26 implied HN points • 10 Mar 26
🎾 Sports Forecasting
  1. COOPER is a new Bayesian college basketball rating system that combines margin-of-victory, opponent strength, pace, and preseason expert polls to produce offensive/defensive (PPPG/PPAG) and Elo-based ratings.
  2. The model changes include separate offensive and defensive ratings, removal of the rule that winners must always gain points, game impact factors that weight close and high-stakes games more, and a time-varying k-factor that updates ratings more aggressively early in the season.
  3. Tournament forecasts combine COOPER with Pomeroy/Her Hoop Stats (COOPER gets 5/8 weight), run conditional (“hot”) simulations that update ratings as simulated games occur, and explicitly model injuries probabilistically to adjust win probabilities.

Congrats To Polymarket, But I Still Think They Were Mispriced

Astral Codex Ten • 16656 implied HN points • 07 Nov 24
🇺🇸 U.S. Politics Forecasting
  1. Polymarket, a prediction market, did well during the recent election by accurately calling results and gaining popularity. However, there are still concerns about accuracy in their pricing, particularly regarding Trump shares.
  2. It's important not to overreact to single significant events when making predictions. Even if things seem to favor one hypothesis after an event, it doesn't mean we should change our overall beliefs drastically.
  3. The reliability of prediction markets like Polymarket compared to non-money forecasting sites like Metaculus is still up for debate. Past performance shows non-money forecasters often have better accuracy, and big bets from individuals can create misleading odds in prediction markets.

Midterm polls have good news, great news, and bad news for Democrats. Wait, what?

Silver Bulletin • 290 implied HN points • 22 Jan 26
🇺🇸 U.S. Politics Forecasting
  1. Current generic-ballot polling shows Democrats up about 5.3 points, which could mean anything from a huge blue wave to a tiny Republican edge once the margin of error is considered.
  2. Historically the party out of the White House tends to win midterms by roughly 4.4 points, so Democrats are positioned to gain seats and could retake the House if that pattern holds.
  3. There’s substantial uncertainty: early-cycle polls have only a moderate correlation with the final House vote and an empirical error of about ±5.7 points, and partisan gerrymandering could blunt Democratic seat gains even with a popular-vote advantage.

🛑🚼 The end of (demographic) history

Faster, Please! • 548 implied HN points • 14 Dec 25
🔬 Science Forecasting
  1. The global fertility transition seems to have largely finished, so the classic story of steadily falling birth rates is no longer the clear master narrative.
  2. Even with that shift, the demographic future is uncertain — demographers don’t know exactly how birth rates, aging, and migration will evolve next.
  3. That uncertainty has big policy and economic implications, because different population paths lead to very different outcomes for growth, labor markets, and public finances.

Open Thread 366

Astral Codex Ten • 7089 implied HN points • 27 Jan 25
🕹 Technology Forecasting
  1. Anyone can share thoughts or ask questions in the open thread. It's a space for discussing anything on your mind.
  2. There are opportunities for people interested in AI safety, including a course that can help you get started in the field.
  3. An AI forecasting project is looking for news outlets to publish articles on future predictions about AI advancements.

FINAL Silver Bulletin 2024 presidential election forecast

Silver Bulletin • 9424 implied HN points • 05 Nov 24
🇺🇸 U.S. Politics Forecasting
  1. In the final simulations for the election, Kamala Harris won slightly more than half of the time, but Donald Trump also had a strong chance, especially with potential ties in the Electoral College.
  2. The modeling approach adjusts for different polling methods and emphasizes reliable polls, aiming to offer a clearer picture of who might win the election.
  3. There’s still uncertainty in elections, and surprising outcomes can happen, so many factors could influence the final results.

The Silver Bulletin Year in Review

Silver Bulletin • 336 implied HN points • 31 Dec 25
🇺🇸 U.S. Politics Forecasting
  1. The newsletter experienced a post-election normalization: overall subscriptions rose about 12% but paid subscribers fell roughly 27%, with realized revenue down less (around 17%) while median post readership stayed strong.
  2. The team is hiring an editor to raise the baseline output while keeping quality high, and plans to diversify beyond politics and sports by launching new forecasting products like a generic ballot, COOPER for college basketball, and continued work on ELWAY/QBERT.
  3. The editorial approach centers on data-driven forecasting and testable predictions, producing notable hits and some misses, and the plan is to be more forward-looking and varied in 2026.

🤖👴 AI vs. age: The great offset

Faster, Please! • 548 implied HN points • 09 Dec 25
💼 Business Forecasting
  1. Many people expect AI to cause a huge economic boom and rapid change across society.
  2. A JPMorgan analysis suggests aging populations will subtract from growth roughly as much as AI can add, so the two forces could cancel each other out.
  3. That means AI might mainly keep economies from shrinking rather than spark a new golden age. So investors and policymakers should temper overly rosy expectations.

So, who’s going to win the Super Bowl?

Silver Bulletin • 280 implied HN points • 07 Jan 26
🎾 Sports Forecasting
  1. This NFL season was unusually chaotic, with most preseason co-favorites failing to make the playoffs and standings flipping because of close-game variance and injuries.
  2. A new predictive model that updates through the season and accounts for quarterback play, injuries, weather, home-field and rest gives refined playoff odds and currently rates Seattle as the likeliest Super Bowl contender, though no team has anywhere near a 50% chance.
  3. Turnover margins are often driven by luck and are less reliable for prediction than efficiency stats like yards per play or sacks, while injuries can meaningfully reduce a team’s rating and playoff prospects.

Analyzing A Critique Of The AI 2027 Timeline Forecasts

Don't Worry About the Vase • 2553 implied HN points • 24 Jun 25
🕹 Technology Forecasting
  1. Critiques are important for improving forecasts. It's good to get feedback and adjust predictions based on detailed analysis.
  2. Modeling progress in AI is tricky and uncertain. It's not easy to predict how quickly AI will advance, and different methods can give very different results.
  3. Forecasts should be communicated clearly, without overly negative language. Clear messaging helps everyone understand the importance and limitations of the predictions.

QBERT NFL Quarterback Ratings

Silver Bulletin • 364 implied HN points • 02 Dec 25
🎾 Sports Forecasting
  1. The QBERT system ranks NFL quarterbacks based on many performance factors, not just traditional stats. It looks at things like rushing yards and how well they handle pressure.
  2. Ratings are adjusted for different conditions, like the strength of the opposing team and the weather, making it fairer across the board.
  3. Along with current ratings, QBERT provides future projections for quarterbacks, taking into account their recent performances, age, and experience.

A Problem of Prophesy

David Friedman’s Substack • 260 implied HN points • 20 Dec 25
🔬 Science Forecasting
  1. Total fertility rate (TFR) is a snapshot-based prediction that can underestimate the number of children women will actually have if they postpone births, while completed fertility rate (CFR) is what determines population change.
  2. There is a biological limit to how late people can have children, so shifting births to older ages can only go so far, though advances in reproductive technology could change that limit.
  3. Life expectancy at birth (an estimated measure) is also a prophecy and can fall during temporary mortality shocks even though completed life expectancy will likely be higher if mortality rates continue to decline.

Try The 2025 ACX/Metaculus Forecasting Contest

Astral Codex Ten • 3854 implied HN points • 20 Jan 25
🔬 Science Forecasting
  1. The 2025 ACX/Metaculus Forecasting Contest is now open for predictions. It's a great opportunity for anyone interested to share their forecasts on various topics.
  2. This year, there are new forecasting bots participating, and it'll be exciting to see how they compare to top human forecasters. The contest wants to explore how well these bots can predict outcomes.
  3. The questions this year are designed to be interesting and relevant, so many people can take part. The contest aims to engage everyone's thoughts on important issues.

20 Predictions for AI in 2026

The Algorithmic Bridge • 286 implied HN points • 12 Dec 25
🕹 Technology Forecasting
  1. A clear set of twenty specific predictions about how AI will develop in 2026 is presented.
  2. The piece reviews results from 2025 predictions and commits to being more specific and accountable to improve forecasting accuracy.
  3. Full access to the detailed content is behind a subscription paywall, though a 7-day free trial is offered.

Who Predicted 2023?

Astral Codex Ten • 8534 implied HN points • 05 Mar 24
🕹 Technology Forecasting
  1. The Annual Forecasting Contest on astralcodexten.com involves participants making predictions about various questions, helping to determine if one identifiable genius or aggregated mathematical predictions work best for foreseeing the future.
  2. The winners of the contest were both amateurs and seasoned forecasting veterans, showcasing a mix of skill and luck in predicting outcomes.
  3. Metaculus outperformed prediction markets, superforecasters, and the wisdom of crowds in the contest, suggesting that consistent high performance might be rare but achievable with specific methods like those used by superforecaster Ezra Karger.

Cogent Analysis

Doomberg • 7086 implied HN points • 07 Feb 24
💰 Finance Forecasting
  1. Analysts focus on continuous learning and understanding, while advocates tend to rationalize or attack inconvenient facts.
  2. Economies heavily reliant on energy resources like Russia may evade recession despite sanctions due to their unique market dynamics.
  3. US economy's short-term resilience and avoidance of recession can be attributed to various energy-related factors, such as LNG export approvals.

Estimating Roche Shipments

ASeq Newsletter • 14 implied HN points • 27 Feb 26
💼 Business Forecasting
  1. You can bound Roche's first-year instrument shipments by comparing to historical first-year shipments of similar sequencing platforms.
  2. Historical examples vary a lot — from roughly 20 units up to about 500 units in their first year — so Roche could plausibly fall anywhere in that range.
  3. Producing a useful estimate will require more data and clear assumptions about market demand, pricing, and manufacturing capacity.

[2025–2026] Week 19: Superforecasting: The Art and Science of Prediction by Philip Tetlock

Behavioral Value Investor • 14 implied HN points • 27 Feb 26
💰 Finance Forecasting
  1. Start building an investing checklist early and update it as your approach evolves so it becomes a reliable repository of your process and decision rules.
  2. Learn and practice forecasting skills by studying what makes superforecasters better than average and by making clear, probabilistic predictions to sharpen judgment.
  3. Share your answers in a single comment and engage with others' responses to learn through feedback and community discussion.

Astera Labs IPO - The Next Connectivity Superhero or Steamrolled By Competition?

SemiAnalysis • 5354 implied HN points • 17 Mar 24
💼 Business Forecasting
  1. Astera Labs is a key player in the AI infrastructure market, supplying connectivity chips to hyperscalers, AI accelerator vendors, and system OEMs.
  2. Connectivity in the datacenter market is highly competitive, but Astera Labs has the opportunity to become a connectivity 'Superhero' by maintaining market share and expanding product lines.
  3. Astera Labs focuses on solving connectivity bottlenecks in high-speed interfaces by offering retimers to improve signal integrity, particularly targeting PCIe 4.0 and 5.0 specs.

The Extinction Tournament

Astral Codex Ten • 9153 implied HN points • 20 Jul 23
🔬 Science Forecasting
  1. Experts and superforecasters had a strong disagreement on the likelihood of global catastrophes.
  2. The tournament explored global disaster risks, with 'Catastrophe' meaning an event killing over 10% of the population, and 'Extinction' meaning reducing human population below 5,000.
  3. The tournament highlighted the challenges in aligning expert predictions, potential biases in forecasts, and the complexities of forecasting AI-related risks.

How to make use of inductive biases

Mindful Modeler • 219 implied HN points • 04 Jun 24
🔬 Science Forecasting
  1. Inductive biases play a crucial role in model robustness, interpretability, and leveraging domain knowledge.
  2. Choosing inherently interpretable models can enhance model understandability by restricting the hypothesis space of the learning algorithm.
  3. By selecting inductive biases that reflect the data-generating process, models can better align with reality and improve performance.

humans vs machines: who's better at predicting 2026? (spoiler: probably neither, we all missed Venezuela)

Five Links (and three graphs) by Auren Hoffman • 56 implied HN points • 15 Jan 26
🕹 Technology Forecasting
  1. A public prediction game pitted humans against three AIs and laid out ten bets for 2026 across health, geopolitics, economy, and AI impact.
  2. The AIs showed very different strategies — ChatGPT was strongly contrarian, Claude hedged cautiously, and Gemini bet optimistically — highlighting divergent machine reasoning.
  3. Both humans and AIs missed a major development in Venezuela, reminding us that experts and models alike can have big blind spots even after modest collective gains in prior years.

Mantic Monday 2/19/24

Astral Codex Ten • 4473 implied HN points • 20 Feb 24
🕹 Technology Forecasting
  1. AI forecasters are becoming more prevalent in prediction markets, with the potential for bots to compete against humans in forecasting events.
  2. FutureSearch.ai is a new company building an AI-based forecaster that prompts itself with various questions to estimate probabilities.
  3. The integration of AI in prediction markets like Polymarket could increase market participation and accuracy, offering a new way to predict outcomes on various topics.

The Bitter Pipeline

Abstraction • 39 implied HN points • 28 Jan 26
🕹 Technology Forecasting
  1. Frontier models scale better than human-designed forecasting pipelines, so the structured process that helped smaller models often adds no value with larger models.
  2. Empirical tests show spending compute on polling and ensembling big models improves forecast skill more than token-heavy steps like classification or decomposition, with ensembling giving measurable uplift while the pipeline did not.
  3. The practical move is to simplify: ensemble aggressively, validate empirically, and keep experimenting with ways to elicit latent model knowledge instead of adding complex hand-crafted processes.