Super Crunchers: Why Thinking-by-Numbers Is the New Way to be Smart

As someone with a background in data science, I found this to be an incredibly inspirational look at the early days of the big data movement. Ayres perfectly captures the friction between 'old world' experts and the new wave of statistical profiling. The discussion on the 100,000 lives campaign was particularly moving; the slogan 'Some is not a number, soon is not a time' should be required reading for every corporate manager. I also loved the quiz on page 113—it really exposes how overconfident we are in our own biased estimates. The truth is, we are all prone to cognitive errors, and this book provides the framework for why we need to let the data do the talking. It’s a pellucid, well-researched argument for computational creativity that has aged surprisingly well despite the rapid changes in technology since its publication.

Ever wonder why a computer can predict a winning Greyhound better than a seasoned track expert? Ayres dives into these types of 'super crunching' scenarios with a style that is both informative and highly readable for non-mathematicians. The Greyhound racing example is particularly mind-blowing because it shows how a simple computer model can yield a 25% profit while human experts consistently fail to keep up with the cold, hard numbers. I found the exploration of how casinos calculate a player's 'pain point' to be both fascinating and slightly terrifying. It’s a wake-up call about how corporations use our own data to manipulate our behavior. Personally, I think the most useful takeaway is that we should always be skeptical of 'expert' intuition when a randomized trial could provide a much clearer answer. This is an excellent, thought-provoking read that encourages you to seek out the data in your own life.

Finally got around to reading this classic, and the discussion on evidence-based medicine alone made it worth the price of admission. The way Ayres explains how Google is transforming medical culture—allowing patients and young doctors to bypass the 'wisdom' of older peers—is still incredibly relevant today. He does a fantastic job of explaining complex concepts like heteroskedasticity and regression without making your head spin. I loved the insight into how T-Rex was reimagined as a scavenger through data; it's a perfect example of why we shouldn't get too attached to our 'intuitions.' This is an inspirational read for anyone who prefers the 'pure truth' of an algorithm over the biased guesses of a human. It’s a bit of a 'money machine' of information. Highly recommended for those who want to understand the shift toward a more quantified society.

Picked this up after reading Freakonomics, and it functions as a solid companion piece for anyone interested in how algorithms are quietly reshaping our everyday lives. Ayres focuses on the shift from expert intuition to data-driven decision-making, using the famous wine vintage prediction formula as a primary example. It’s fascinating to see how simple variables like rainfall and temperature can outperform the palates of world-renowned critics. While the book was written in 2007, the core message about the power of regression remains incredibly relevant today. Some sections feel slightly repetitive, and the author definitely likes to pat himself on the back, but the insights into evidence-based medicine and the '100,000 lives campaign' make it a worthwhile read. To be fair, if you already work in big data, some of this will feel like an introductory text, but for the general public, it's an eye-opener.

After hearing so much about 'computational creativity,' I decided to give this a go and wasn't disappointed by the breadth of real-world examples provided. Ayres makes a compelling case for why we should be wary when corporations start being 'nice' to us; it usually means they've crunched the numbers and realized we're overpaying. The section on car dealership discrimination was a gut-punch, showing how statistical profiling can be used to exploit certain demographics. It's an analytical look at the shift of power from the periphery to the 'Super Crunching' center. The writing is accessible, though the author's tone can be a bit self-congratulatory at times. Still, the explanation of neural networks versus traditional regression was one of the clearest I've ever read. It’s a great reference for anyone who wants to understand the logic behind the algorithms that run our world.

In my experience, books about math are either too dense or too shallow, but Ayres finds a decent middle ground here. He highlights how randomization and large sample sizes are the only real ways to determine what actually works in government and medicine. The bit about how annual physical exams are largely obsolete according to data was shocking but well-supported. It really opens your eyes to the predictability of the world. While the 2007 publication date makes some of the technology talk feel like ancient history, the underlying logic of data-driven decision-making is timeless. I particularly appreciated the discussion on how the internet is disrupting the traditional 'sage' status of doctors. It's a well-written, informative text that will definitely make you rethink your own decision-making process. I’m glad I didn't shove it back on the shelf.

Look, the book is a bit dated now, especially when Ayres discusses the 'novelty' of internet data mining which is now just standard corporate practice for almost every company on earth. I found the chapter on regression to the mean interesting, particularly the bit about how tall fathers often have shorter sons, but I wish he had snuck those statistical foundations in much earlier. The Greyhound racing study mentioned in the beginning felt like a missed opportunity; if a computer model is a 'money machine' with 25% profit, why isn't everyone doing it? Something felt a little fishy there. It’s a decent introductory text if you’ve never thought about A/B testing or randomization before. However, if you’ve read anything published in the last five years on big data, you can probably just skim this one. It's okay, but not groundbreaking anymore.

This book attempts to bridge the gap between human intuition and mathematical algorithms, though it sometimes leans too far into the 'formula is always right' camp. I struggled with the idea that 'direct instruction' for teachers—literally reading from a script—is the best way to educate children just because the numbers look good on a chart. Statistics are powerful, but they don't always capture the nuance of the human experience. The wine vintage discussion was the highlight for me, as it perfectly illustrated the outrage experts feel when their 'wisdom' is challenged by a simple equation. It's a bit of a mixed bag. Some chapters are brilliant and others feel like filler seasoned with dabs of fun. It’s worth reading for the parts on standard deviation, but be prepared for some 'blah' stretches in the middle.

The word 'supercrunch' appears so many times in these 270 pages that it starts to lose all meaning. Dr. Ayres clearly wanted to write the next big pop-economics hit, but he missed the mark by being overly repetitive and a bit too self-promotional regarding his ties to Steven Levitt. It’s just grating. The book focuses heavily on anecdotes rather than teaching the reader the actual statistical techniques he champions. I was looking for a deeper dive into how to apply these methods, yet I walked away only knowing that regression was discovered by Darwin’s cousin. Frankly, it feels like a collection of research descriptions fleshed out with too much filler. If you want to know that computers are better than people, this book tells you that a billion times. It doesn't really show you how to join the revolution yourself.

Not gonna lie, I expected a lot more 'how-to' and a lot less 'wow, look at me.' Ayres spends so much time using the prefix 'nano-' and the word 'supercrunch' as every possible part of speech that the actual message gets buried. He stresses the need for the public to understand statistics but then fails to go over any actual techniques in a meaningful way. It feels very much like a Freakonomics wannabe that lacks the charm and tight editing of the original. The anecdotes vacillate between being creepy and overly exuberant. For example, the way he describes data mining in dating and medicine feels a bit paranoid at times. It’s a short enough book that it doesn't completely waste your time, but it certainly didn't leave me feeling like I learned anything I couldn't have found in a long-form magazine article. Very repetitive.