The hottest Forecasting Substack posts right now

Enrollment in SFUSD is down by almost 4,000 students (7.6%) since before the pandemic.
Charter schools overall saw a 5.8% decrease in enrollment, with variations among individual schools.
Cohort survival rates have recovered since the pandemic, but future enrollment projections for SFUSD are challenging due to declining birth rates and migration trends.

Exploring counterfactual scenarios helps forecast future trends by imagining a world without specific factors like large language models (LLMs).
Using the "outside view" involves making predictions based on broader trends and historical data rather than focusing on specific instances.
Monte Carlo simulations provide an empirically-grounded view by generating random future scenarios based on historical changes, aiding in predicting potential outcomes.

Consider the reasons behind choosing programming languages like Go over Rust, and how they impact problem-solving strategies.
Explore innovative approaches like decoder-only foundation models for time-series forecasting to achieve high performance with less complexity.
Reflect on the impact of intentional choices on software development, understanding tools like TLA+ for formal method and the importance of thoughtful deployment automation.

Wall Street analysts have consistently missed S&P 500 year-end targets by an average of 15.7% from 2018 to 2023.
It's hard for even the most renowned financial firms to predict exact stock market values, showing the importance of personal research.
Despite sophisticated analysis, Wall Street analysts often get S&P 500 projections wrong, emphasizing the value of independent thinking in investment decisions.

Graph databases and graph analytics are growing in interest, with use cases and applications expanding.
The NLP Summit offers insights from leading organizations and researchers in the field of Natural Language Processing.
Tools like Darts for time series forecasting and River for online machine learning are open-source libraries enabling easier adoption of advanced machine learning techniques.

Get a weekly roundup of the best Substack posts, by hacker news affinity:

The correlation between weather data like solar radiation and solar energy with solar production is high, indicating a predictive relationship.
By using historical and forecasted weather data, it's possible to project solar energy production up to two weeks in advance, offering insights for planning.
Accuracy of solar energy predictions from sources like Visual Crossing is crucial for reliable projected energy production outcomes.

Expectations are key to economic models because they shape how people behave and react to changes in the economy. For example, if people expect prices to rise, they may ask for higher wages.
There is confusion about whether expectations tend to overreact or underreact to information. Evidence shows that expectations can do both—people might overreact to recent events but underreact to larger economic trends.
Bias in expectations is often studied, but noise—random fluctuations and errors—is just as important and can affect forecasts significantly. Understanding both can help improve how we predict economic outcomes.

Complex models can struggle when predicting unpredictable human behavior. Simple models might work better in uncertain situations.
Small changes in a complex model can lead to large unexpected outcomes, a phenomenon known as the butterfly effect.
When faced with uncertainty, it's better to simplify models by focusing on key variables and reducing complexity.