Questions for you:
- When did you last describe something as “a bit random” — and did you actually do the maths, or did you just reach for a convenient explanation?
- Can you think of a recent coincidence that felt significant? What would the actual probability calculation look like if you tried to do it properly?
- How often do you notice the connections that don’t happen, compared with the ones that do?
Organisational applications:
Auditing the hidden structure in “surprising” outcomes: Organisations regularly treat unexpected results — an unusually high sales month, a product that suddenly finds a new audience, a partnership that emerges from a chance meeting — as evidence of something meaningful: strategy working, or luck favouring the bold.
Before drawing conclusions, it’s worth asking whether the result was genuinely improbable or, like the Prague penguins photograph, the product of a curated pool and a large number of trials. A sales spike after a product launch isn’t surprising if you launched to a warm audience. A chance meeting at a conference isn’t surprising if you attend six conferences a year. Organisations that mistake the predictable for the serendipitous make poor decisions about what to repeat and what to attribute to skill. Build a habit of asking: how many opportunities were there for this outcome to occur, and how similar were the conditions to what was needed?
The base rate problem in performance assessment: The page’s core point — that we reach for faith or narrative when the maths would give a cleaner answer — applies directly to how organisations evaluate people and teams. When a manager consistently delivers good results, it’s tempting to conclude they are simply excellent.
But performance assessments rarely account for base rates: how often does anyone in this role, in this market, in these conditions, deliver comparable results? A fund manager who beats the index for three years running sounds impressive until you note that, given the number of fund managers operating, some are statistically near-certain to do so by chance. This isn’t an argument for ignoring performance data — it’s an argument for situating it within a realistic probability framework before promoting, rewarding, or replicating it.
Designing systems that don’t mistake noise for signal: The photo randomiser in the story wasn’t random at all — it drew from a pool specifically chosen to resonate with a particular audience. Many organisational data systems share a common hidden structure: customer feedback surveys that reach only satisfied customers, employee engagement scores that capture only those willing to complete surveys, A/B tests run on audiences already predisposed toward one variant.
The results from these systems feel like meaningful signals because they’re presented as data, but they’re shaped by selection in ways that make coincidence and apparent confirmation almost inevitable. Periodically auditing the construction of data sources — asking who is and isn’t in the pool, and what that means for the apparent result — is more valuable than debating the results themselves.
Further reading
On probability intuition and why it fails us:
The Improbability Principle: Why Coincidences, Miracles and Rare Events Happen Every Day by David J. Hand. Hand sets out five laws — including the law of truly large numbers and the law of selection — that explain why apparently miraculous coincidences are statistically inevitable. Directly relevant to the card’s argument.
https://timharford.com/books/worldaddup/ by Tim Harford. A practical guide to statistical thinking with particular attention to how hidden structure in data produces misleading apparent coincidences and surprises.
On coincidence and the limits of intuition:
The Drunkard’s Walk: How Randomness Rules Our Lives by Leonard Mlodinow. Covers the psychological mechanisms by which we systematically overestimate the significance of random events, including the clustering illusion and our tendency to construct narrative explanations for chance outcomes.
Innumeracy: Mathematical Illiteracy and Its Consequences by John Allen Paulos (Penguin, 1990). The foundational text on why coincidences feel so much more remarkable than they are — Paulos works through the birthday problem and similar examples to show how systematically we underestimate the frequency of chance matches.
On base rates and performance assessment:
Thinking, Fast and Slow by Daniel Kahneman (Allen Lane, 2011). Kahneman’s treatment of base rate neglect — our tendency to ignore background frequency when assessing specific cases — is the clearest account of why the Susi coincidence feels spooky even when it isn’t.
The Success Equation: Untangling Skill and Luck in Business, Sports, and Investing by Michael Mauboussin (Harvard Business Review Press, 2012). A rigorous attempt to separate skill from luck in performance data, with practical methods for adjusting assessments to account for the base rate of success in a given domain.
About the image
As explained in the story, these are three of the delightful penguins from The Cracking Art Group in Kampa Park, Prague.
Photo montage and photo by Matt Ballantine, 2026
Random the Book 