What I'm Reading This Week (2024/10.27-11.02)

Table of Contents

• Trends
• What I’m Reading This Week (2024/10.27-11.02)
  • 1. Prediction markets and the need for “dumb money” as well as “smart money”
  • 2. A Collection of Useful Resources for Data Visualization
  • 3. An Introduction to Business Analytics for Business Analysts
  • 4. Crash Course: Probability
  • 5. Veridical Data Science: The Practice of Responsible Data Analysis and Decision-Making

Trends

No trends this week

What I’m Reading This Week (2024/10.27-11.02)

Good morning, this is the first week of November 2024.

It has been a disappointing and disheartening week. In the early hours of the weekend, I witnessed China’s League of Legends team, BLG, squander two opportunities in the global finals, ultimately losing to the T1 team led by Faker.

Thirteen years have passed, and now in the 14th season, I no longer have the mood to discuss who is to blame. As the lifespan of this game nears its end, I have yet to see a team composed entirely of Chinese players win the championship.

1. Prediction markets and the need for “dumb money” as well as “smart money”

• Prediction markets are accurate because of “smart money” that makes logical decisions, which can correct discrepancies between current odds and available information.
• “Dumb money” comes from individuals participating in the market for entertainment or passive investment purposes, and its presence gives confidence to “smart money” participants about the fairness of the market.
• The effectiveness of prediction markets is limited due to a lack of sufficient “dumb money.”
• Results from aggregated predictions of a broad audience generally surpass individual predictions.
• Gambling markets allow participants to bet based on private knowledge, helping to quickly adjust prices in response to new information.
• Both “smart money” and “dumb money” are essential for the effective operation of markets; without “dumb money,” the incentive for “smart money” to correct mispricing decreases.
• Non-sports prediction markets are extremely rare, and even in countries where they are legal, they remain unpopular.
• Participants in prediction markets can be categorized as savers, gamblers, and experts; however, without savers and gamblers, market efficiency would be impacted.
• Prediction markets are zero-sum games, lacking the long-term growth potential of other investment markets, which discourages savers from participating.
• Funding prediction markets could enhance their effectiveness, although implementing such funding presents practical challenges.
• The current demand for future predictions is met by existing markets and companies, raising questions about the necessity and viability of prediction markets. Here’s the translation in English, with formatting preserved:

2. A Collection of Useful Resources for Data Visualization

3. An Introduction to Business Analytics for Business Analysts

• A guide for beginners, aimed at developing data-driven business analysts.
• Focuses on building essential skills to enable analysts to combine real-world narratives with mathematical concepts and code, particularly in R programming and Bayesian inference.
• Uses the tidyverse and causact packages, emphasizing data manipulation (dplyr) and data visualization (ggplot2). The goal is to help analysts tell compelling stories with data and make decisions through a unified narrative, math, and code framework.
• Introduces Bayesian inference through graphical models and Directed Acyclic Graphs (DAGs), using the causact package integrated with Python’s NumPyro for efficient Bayesian computation. Readers will learn how to use these tools for parameter estimation, predictive checks, and strategy development based on statistical data.
• Each chapter is accompanied by corresponding YouTube videos.
• The Capstone Project transforms a business problem into an analytical model, enabling data-driven decision-making and effective communication of analytical insights.
• The datasets used in the book are free, and there are cloud computing options available.

4. Crash Course: Probability

5. Veridical Data Science: The Practice of Responsible Data Analysis and Decision-Making

• This is an open-source book on reliable data science, to be published by MIT Press at the end of 2024.
• Starts from ambiguous domain problems and messy data. This new open-access text explores the Veridical framework and demonstrates its application through real-world case studies.
• The book is primarily divided into three parts:
  • Part 1 introduces core concepts in reliable data science.
  • Part 2 focuses on data preparation, exploration, and description.
  • Part 3 concentrates on predictive problems.
• Central to the book is the PCS framework: Predictability, Computability, and Stability.
• Suitable for readers with a basic knowledge of calculus and linear algebra.
• Provides supporting code in both R and Python.