Questions for you:
- For the roles in your organisation where you most heavily weigh track record, how much randomness is involved in that domain? Are you looking at twenty years of root canals or twenty years of fund management?
- Can you think of someone you or your organisation hired or commissioned primarily on the basis of past success in a high-variance field? How confident are you that the success reflected skill rather than favourable conditions?
- What would change about your hiring or procurement processes if you explicitly accounted for the randomness level of the domain rather than treating all track records as equally informative? Who might you hire instead? How much money might you save?
Organisational applications:
Matching the weight given to track records to the randomness of the domain: The story’s central practical instruction is explicit: in low-variance domains, past performance is a reliable signal and should be weighted heavily; in high-variance domains, it is a much weaker signal and should be weighted accordingly.
The problem is that most organisational hiring and commissioning processes do not make this distinction — they apply the same track-record-focused approach regardless of whether the work is closer to dentistry or to fund management. The immediate practical step is to assess the domain’s variance before deciding how much weight to give historical outcomes. In high-variance domains, this means shifting weight towards other evidence: the quality of reasoning in the candidate’s explanations, the process behind their decisions, and references that speak to how they operated rather than just what they achieved.
The “just world hypothesis” in recruitment: The story names the “just world hypothesis” — the belief that success is always earned and the scoreboard can be trusted. This belief is pervasive in recruitment contexts, where impressive outcomes are routinely taken as direct evidence of impressive capability. It is particularly resistant to challenge because it flatters successful candidates and reassures the people selecting them.
Building explicit friction into high-stakes hiring decisions in high-variance domains — requiring candidates to explain the conditions behind their successes as well as the successes themselves, or asking structured questions about decisions that did not go well — is a partial corrective, though it requires a culture that does not treat any scepticism about track records as inappropriate.
Distinguishing the dentist from the dice in procurement and advisory selection: The same logic applies to commissioning external expertise. A consulting firm with an impressive client list may have delivered genuine value, or may have been engaged during favourable market conditions that made most interventions look successful. A fund manager with a five-year record of outperformance may be skilled or may have been running a strategy that happened to suit the market cycle.
Before commissioning on the basis of track record, the relevant question is: in this domain, over this period, what proportion of similarly positioned practitioners also performed well? If the answer is “most of them,” the track record is much less informative than it appears. If the answer is “very few,” it becomes more meaningful — though still not conclusive.
Further reading
On track records, skill, and the randomness of domains:
The Success Equation: Untangling Skill and Luck in Business, Sports, and Investing by Michael Mauboussin. The most systematic available treatment of how to assess the skill-luck balance in different domains, with direct guidance on what this implies for how much weight to give historical performance in each.
Fooled by Randomness: The Hidden Role of Chance in Life and in the Markets by Nassim Nicholas Taleb. Taleb’s account of how traders and professionals mistake lucky runs for skill is the sharpest available critique of track-record-based hiring in high-variance domains, with sustained attention to the fund management examples the story references.
On structured assessment and reducing outcome bias in evaluation:
Thinking in Bets: Making Smarter Decisions When You Don’t Have All the Facts by Annie Duke. Duke’s framework for separating decision quality from outcome quality is directly applicable to how interviews and reference checks should be designed in high-variance domains: the focus should be on process and reasoning, not results.
Noise: A Flaw in Human Judgement by Daniel Kahneman, Olivier Sibony and Cass R. Sunstein. The chapters on hiring and structured assessment cover the evidence that unstructured interviews weighted toward track records produce systematically biased and unreliable outcomes, with concrete suggestions for more robust approaches.
On base rates, market conditions, and contextualising performance:
Superforecasting: The Art and Science of Prediction by Philip Tetlock and Dan Gardner. Tetlock’s argument for always establishing the base rate before evaluating individual performance is directly relevant: the question is not whether someone’s track record looks impressive in isolation but whether it looks impressive given what most people in that position achieved over the same period.
Against the Gods: The Remarkable Story of Risk by Peter L. Bernstein. Bernstein’s history of risk and probability provides the context for understanding why different domains sit at different points on the skill-luck spectrum, and why the financial domains the story singles out are particularly resistant to track-record-based prediction.
About the image
In the summer of 2025, a project team I was working with concluded their engagement. As a celebration, we went to a particularly nice pub in Notting Hill and had lunch together. This was my plate – a particularly nice serving of ham, egg and chips. The egg was well-cooked.
Photo montage and photo by Matt Ballantine, 2026
Random the Book 