Ian Ayres’s Super Crunchers is an enthusiastic tour through the growing power of statistical analysis and large-scale data processing to make decisions that were once the exclusive domain of human experts. Published in 2007, the book argues that algorithmic and regression-based approaches are quietly outperforming doctors, judges, wine experts, financial analysts, and a host of other credentialed professionals in task after task. Ayres, a law professor and economist at Yale, writes with the energy of a true convert, peppering his argument with surprising case studies and accessible explanations of how techniques like randomized controlled trials, logistic regression, and neural networks work in practice.

The book’s central claim is not merely that computers are fast, but that large datasets reveal hidden patterns and correlations that even the most experienced human intuition consistently misses or distorts. Ayres traces the rise of what he calls “super crunching” — the mining of massive databases to generate predictions and decisions — across fields as varied as baseball (the sabermetrics revolution made famous by Moneyball), medicine, marketing, and government policy. He is careful to acknowledge the legitimate role that human judgment still plays, particularly in framing the right questions and interpreting results, but his overall arc is one of confident advocacy: institutions and individuals that embrace data-driven decision-making will outcompete those that cling to intuition and tradition.

Ayres writes for a general audience and keeps the mathematics light, favoring vivid anecdotes over technical formalism. His tone is conversational and occasionally playful, and he clearly enjoys revealing the counterintuitive conclusions that emerge when gut feelings are tested against evidence. At the same time, he grapples — if somewhat briefly — with the ethical and social implications of living in a world where algorithms increasingly sort, rank, and decide on human lives, raising questions about transparency, fairness, and the erosion of professional autonomy.

Key takeaways

• Algorithms routinely beat experts. Drawing on decades of research by psychologist Paul Meehl and others, Ayres shows that simple statistical models outperform clinical judgment in predicting everything from psychiatric recidivism to wine vintage quality, challenging the assumption that credentialed expertise is irreplaceable.

• Randomized experiments are the gold standard — and they are spreading. Companies like Amazon, Capital One, and Harrah’s run thousands of randomized controlled trials on their customers to optimize pricing, interfaces, and offers, borrowing the logic of clinical drug trials and applying it at commercial scale.

• Regression analysis uncovers non-obvious relationships. Orley Ashenfelter’s famous equation for predicting Bordeaux wine prices from rainfall and temperature data is used as a recurring example of how a few variables, rigorously analyzed, can match or surpass the judgment of the most respected sommeliers and critics.

• Data-driven decision-making is spreading into deeply personal domains. From sentencing guidelines in criminal courts to personalized medicine and targeted political advertising, super crunching is reshaping decisions that people once considered too nuanced or human for quantitative treatment.

• The “human in the loop” still matters, but in a changed role. Ayres argues that the most effective systems combine algorithmic output with human oversight — not to second-guess the statistics, but to ask better questions, catch data errors, and handle genuinely novel situations outside the model’s training.

• Resistance from incumbent experts is predictable but costly. Professionals whose status and income depend on the perceived irreplaceability of their judgment have strong incentives to dismiss statistical approaches, and Ayres documents cases — in medicine especially — where this resistance has delayed better outcomes.

• Transparency and accountability in algorithms are emerging concerns. As super crunching moves from back-office analytics to decisions that directly affect individuals’ freedom, credit, and health, Ayres flags the need for scrutiny of how models are built, what data they use, and whether their outputs can be meaningfully explained or contested.