You know the story. You admire its elegance. You have probably retold it at a dinner party. And yet, today — almost certainly this very hour — you are repeating the exact error it was designed to prevent. This is an essay about why that happens, and what, if anything, can be done about it.
I. A Hungarian Mathematician Steps Into the Room
It is 1943. The United States Army Air Forces are losing bombers over occupied Europe at a rate that is becoming strategically untenable. The planes return — or rather, some of them do — pockmarked with bullet holes clustered along the fuselage, the wings, the tail sections. The cockpits and engines, by contrast, appear relatively unscathed. The obvious conclusion, the one reached by men of evident competence and reasonable intelligence, is to reinforce the areas under fire: bolt extra plating to the wings, add armour to the fuselage, protect the places that the data show are most vulnerable.
Then Abraham Wald walks in.
Wald is a Hungarian-Jewish mathematician who fled Austria in 1938 after the Nazi annexation — an act of exile that saved his life, given that almost his entire family would perish in the Holocaust. By 1943 he is a member of the Statistical Research Group at Columbia University, a wartime body tasked with applying mathematical rigour to military problems. He is so trusted, and his work so sensitive, that his secretaries had to physically remove notepaper from his hands the moment he finished writing — he lacked the security clearance to read back his own classified analyses. As Jordan Ellenberg recounts in How Not to Be Wrong: The Power of Mathematical Thinking (2014), there was a running joke in the Statistical Research Group about Wald's Kafkaesque situation: the man solving the problem was not permitted to know the solution.
What Wald tells the officers is this: armour the engines. Armour the cockpit. Reinforce precisely the areas showing the least damage.
The reasoning is as simple as it is devastating. The planes they are examining are the ones that came back. The bullet holes visible on those fuselages are not a map of danger; they are a map of survivability. The aircraft hit in the engines and the cockpit are not on the tarmac because they never returned. They are scattered across fields in France and Germany, unreachable and uncounted. The data set before the officers is not a sample of all planes hit by enemy fire. It is a sample, systematically biassed, of planes resilient enough to absorb damage in non-critical zones and still limp home.
The absence of bullet holes on the engines is not a sign that the engines were protected. It is a sign that planes struck in the engines did not survive to be counted.
Wald's original 1943 memorandum, "A Method of Estimating Plane Vulnerability Based on Damage of Survivors," was classified and circulated only internally. It was not published in the open literature until 1981, when the Center for Naval Analyses released it. By then, the insight it contained had migrated from military statistics into economics, finance, medicine, management theory, and psychology — becoming, arguably, one of the most consequential statistical ideas of the twentieth century.
And it has changed almost nothing about the way most people reason.
II. The Bias with a Name (That Changes Nothing)
In cognitive science, naming a bias is usually considered the first step towards correcting it. The existence of a word — survivorship bias, availability heuristic, confirmation bias — implies the possibility of inoculation. You learn the concept, you add it to your mental toolkit, and thereafter you catch yourself before falling prey to it. This is the standard pedagogy of popular rationalism, the premise underlying everything from Daniel Kahneman's Thinking, Fast and Slow (2011) to the vast ecosystem of cognitive-bias explainers that populate the internet.
The difficulty is that survivorship bias does not work this way. It is stubbornly resistant to the name.
The reason lies in the architecture of the error itself. Survivorship bias is not merely a failure of calculation. It is a failure of imagination — specifically, the failure to conjure into existence the things that are not there. And imagination of absence is cognitively far more demanding than perception of presence. The bullet holes are visible. The missing planes are not. The successful fund is in the brochure. The liquidated fund is not even a memory.
Kahneman and his long-time collaborator Amos Tversky mapped this territory across decades of experimental work. Their concept of the availability heuristic, introduced in a landmark 1973 paper in Cognitive Psychology, describes a mental shortcut in which people judge the frequency or probability of events based on how easily examples come to mind. Success stories are vivid, widely repeated, emotionally engaging, and therefore available. They dominate the mental inventory from which we draw our probabilistic intuitions. Failure stories are quiet, embarrassing, commercially unremarkable, and rapidly forgotten.
Tversky and Kahneman demonstrated this with deceptive elegance: when experimental participants were read lists containing either nineteen famous women and twenty less-famous men, or nineteen famous men and twenty less-famous women, the majority incorrectly identified the gender with the famous names as the more numerous, even when the actual count favoured the opposite. Fame — a proxy for survival in the attention economy — warped numerical judgement entirely.
Layered on top of the availability heuristic is what Kahneman called the narrative fallacy: the human compulsion to construct causal stories from sequences of events, regardless of whether causation is actually present. When we hear that a tech founder dropped out of university and later built a company worth billions, our brains do not register the event as a data point in a large distribution of outcomes. They construct a story: the dropout caused the success. The unconventionality generated the result. We then carry this story into our own decisions as a template for action, applying a lesson extracted from a sample of one.
Nassim Nicholas Taleb, whose The Black Swan (2007) is perhaps the most sustained literary engagement with the problem of invisible evidence, coined the term "silent evidence" for the data systematically excluded from our view by the mechanics of selection. He illustrates it with an ancient story: someone shows the philosopher Diagoras of Melos a series of painted portraits of shipwreck survivors who had prayed to the gods before the storm. The implication is that prayer works. Diagoras asks the obvious question that the paintings do not answer: where are the portraits of those who prayed and drowned? The graveyard of the faithful is silent. It always is.
III. The Museum of Survivors: Finance
No domain has been more thoroughly distorted by survivorship bias than investment — and no domain has had it more rigorously documented.
The academic literature on the subject dates back to at least 1992, when Stephen J. Brown, William Goetzmann, Roger G. Ibbotson, and Stephen A. Ross published "Survivorship Bias in Performance Studies" in the Review of Financial Studies. They demonstrated mathematically that if mutual fund data is drawn only from funds currently in operation — which is what happens when you consult a living database — the apparent evidence that past performance predicts future performance is itself an artefact of survivorship bias. Funds that perform poorly close. Closed funds exit the database. The database then shows only the winners. The winners appear to be the rule. They are not.
Burton Malkiel, in a related analysis, found that adjusting for defunct funds reduced apparent returns by approximately 1.5 percentage points annually. A Savant Capital Management study examining Morningstar mutual fund data for the period 1995 to 2004 found that dropping defunct funds from the Morningstar database boosted reported returns on average by 1.3 per cent per year over the full ten-year period. More recently, the 2025 analysis of the NIFTY Smallcap 250 index by Rishi Shukla et al., published as an arXiv preprint, found that survivorship bias inflated apparent index returns by between 2 and 5 percentage points annually in emerging market small-caps, a context where attrition rates are highest and the bias is therefore most severe.
What does this mean in practice? It means that virtually every "ten-year performance" figure you have ever seen in a financial prospectus, a wealth management advertisement, or a retirement planning document has been calculated on a data set that excludes the experiments that failed. You are being shown the planes that came back, presented as though they represent the entire fleet.
The venture capital world compounds the problem with a rhetorical strategy that has become almost universal: the portfolio boast. A VC firm names its successes — the company acquired for nine figures, the IPO that returned forty times the investment — without mentioning the twenty startups that went to zero. This is not dishonesty in the narrow legal sense; it is the structural consequence of a world in which success is visible and failure disappears. The firm genuinely believes its track record reflects skill. In many individual cases, it may. The difficulty is that there is no statistical method by which an investor hearing the boast can distinguish skill from the operation of survivorship bias on a portfolio of longshots.
Ellenberg, in How Not to Be Wrong, illustrates this with what he calls the Baltimore stockbroker thought experiment. Imagine receiving an unsolicited letter predicting that a particular share will rise next week. It rises. The following week, another letter arrives with another prediction. It is correct again. By the tenth consecutive correct prediction, you are reaching for your chequebook. What you do not know is that the letter was sent to 10,240 people in week one, with half receiving the bullish prediction and half the bearish. Each week, the half who received the wrong prediction are dropped from the list. After ten weeks, ten people remain, each of whom has received ten consecutive accurate predictions from what they are now certain is an oracle. There was no oracle. There was a systematic elimination of the evidence of error.
IV. The Strategy Shelf: Management and Business
The business section of any well-stocked bookshop is, in one sense, the most concentrated physical repository of survivorship bias in human civilisation. Book after book analyses companies that have thrived — the ones that "made it home" — and extracts from their behaviour a set of principles it invites the reader to apply. The companies that followed identical or similar principles and collapsed are not examined. They do not get chapters. They barely get footnotes.
Jim Collins and Jerry I. Porras's Built to Last (1994) and Collins's subsequent Good to Great (2001) are the most prominent examples of what one might call the survivor-literature genre. Both books identify characteristics shared by a group of exceptional, long-lived companies, and argue that these characteristics explain the companies' exceptional performance. The methodological problem, identified by Phil Rosenzweig in The Halo Effect (2007) and various subsequent critics, is that the books study only the successful companies. The analytical question — do these characteristics predict success, or are they simply common features of all large, long-running corporations, including those that subsequently failed? — cannot be answered from the data Collins and Porras assembled, because they did not assemble the data on the companies that did not make the list.
Rosenzweig documented what he called the halo effect: the tendency to attribute the characteristics of successful companies to causal explanations that are themselves derived from the success. A company that is performing well is described by analysts, journalists, and former employees as having strong leadership, an excellent culture, a coherent strategy, and decisive management. The same company, five years later, after a catastrophic product failure, is described as having been poorly led, culturally complacent, strategically confused, and managerially arrogant. The characteristics did not change. The financial performance changed, and our perception of the characteristics followed the performance retroactively.
The startup ecosystem has developed its own particularly pure form of the problem. The "founder playbook" — the set of behaviours attributed to successful entrepreneurs and endlessly prescribed in accelerator programmes, business school case studies, and the advice columns of technology publications — is assembled almost entirely from observation of survivors. "Move fast and break things." "Ignore the naysayers." "Focus on growth above profitability." "Hire for culture fit." Each of these prescriptions has been extracted from the behaviour of companies that succeeded. Whether it contributed to their success, was irrelevant to it, or was in fact a liability that they survived despite rather than because of, cannot be determined from the sample, because the companies that moved fast and broke things and subsequently broke themselves are not in the data.
This is not a criticism of individual entrepreneurs or business books. It is a description of an epistemological condition: we cannot learn from the failures we cannot see.
V. The Silent Graveyard: Personal Life and Career
The survivorship bias of business literature is, at least, commercially motivated. Less excusable — or perhaps simply more poignant — is its operation in the personal domain, where the stakes are a human life spent rather than a portfolio misallocated.
Career advice presents an especially fertile field for the bias. The testimony on which most careers advice rests is gathered from the people who made it: the professor emeritus who counsels students to pursue their passion regardless of the job market, the serial entrepreneur invited back to campus to describe the virtues of risk-taking, the artist who followed her calling and eventually found recognition. These testimonies are not false. They are simply drawn from a sample that systematically excludes the people who followed the same advice and arrived at different outcomes — who pursued the passion and found it unremunerated, who took the risk and went bankrupt, who followed the calling into poverty or obscurity.
The psychologist Barry Schwartz, in The Paradox of Choice (2004), notes that instructions to "follow your passion" carry an implicit and unexamined premise: that the passion exists prior to the career, that it can be identified in advance, and that it will remain motivating under the grinding conditions of professional life. Cal Newport, in So Good They Cannot Ignore You (2012), marshals evidence that the causal arrow often runs the opposite direction: career satisfaction tends to precede the subjective sense of passion, not follow it. People become passionate about things they are good at, and they become good at things through deliberate engagement over time. The "passion first" advice, distilled from the retrospective testimony of successful people who naturally frame their career as the pursuit of a pre-existing calling, may be describing an effect while presenting it as a cause.
Relationship advice is subject to an even more radical form of the problem, one that has barely been studied. The couples who appear in magazine features about long marriages, who describe their secrets to decades together, who speak at workshops on sustaining partnerships — these are, by definition, the unions that survived. The divorces do not host podcasts. The separations do not grant interviews. We receive instruction in marriage from those who have not failed at it, without any mechanism for comparing their traits, behaviours, and circumstances against those of people who followed similar principles and arrived at dissolution. This is not a knock against long-married couples or the genuineness of their experience. It is an observation about the systematic incompleteness of any testimony drawn only from those who arrived.
The education system participates in this distortion with particular vigour. The people invited to advise on the pathway to academic or professional success are overwhelmingly those who succeeded within the existing system. The students for whom the system did not work — who left, who were failed, who spent years in environments misaligned with their learning or who were disadvantaged by the structural biases embedded in selection processes — are not convened to assess the system's efficacy. The advice that flows from top-down mentoring programmes, alumni networks, and commencement addresses is advice given by survivors to aspiring survivors, with no representation from those the system left behind.
VI. The Bias in the Lab: Medicine and Science
If the human cost of survivorship bias in career and relationship decisions is dispersed and difficult to quantify, its cost in medicine is both direct and measurable.
Clinical research is vulnerable to the bias at multiple levels. The most widely discussed is publication bias: the systematic tendency of journals to publish studies with statistically significant, positive results while declining or ignoring studies that find no effect, or negative effects. The consequence, well-documented since at least the 1990s, is that the published medical literature overestimates the efficacy of treatments relative to the truth. A meta-analysis conducted on the published literature is, in effect, a study conducted on a sample of results biassed towards the positive — which is to say, a sample truncated by the same logic as Wald's fleet. The studies that "survived" the publication process are the ones that returned from the mission. The ones that found no effect crashed silently.
This is not merely a theoretical concern. A 2014 analysis published in PLOS Medicine examined the publication rates of clinical trials registered at ClinicalTrials.gov and found that a substantial fraction of completed trials are never published, with negative results facing significantly higher rates of non-publication than positive results. The result is a medical literature that presents physicians and patients with a systematically optimistic picture of therapeutic efficacy.
Within individual trials, survivorship bias operates through the mechanism of patient dropout. If patients who experience adverse effects or fail to respond to a treatment withdraw from a study, and the final analysis is conducted only on those who completed the trial, the estimated efficacy of the treatment will be inflated by the exclusion of the non-responders. The standard methodological corrective is the intention-to-treat analysis, in which all patients originally randomised to a treatment arm are included in the final analysis regardless of whether they completed the study — but this requires that investigators collect data on dropouts, which is itself a logistically demanding and frequently incomplete process.
The physicist and risk analyst Taleb, extending his analysis in Antifragile (2012), notes that evidence-based medicine is itself susceptible to a second-order survivorship bias: the treatments and interventions that get studied, and therefore generate the evidence base, tend to be those that are patentable, marketable, and amenable to the randomised controlled trial format. Lifestyle interventions, dietary changes, and long-term behavioural modifications are systematically under-studied because they are not commercially capturable and because the relevant trials are logistically prohibitive. The evidence base is therefore biassed not only towards positive results but towards the class of interventions that can be studied — which is not the same as the class of interventions that work.
VII. The Deep Mechanism: Why the Story Fails to Stick
It is worth dwelling on the paradox that sits at the heart of this problem. Survivorship bias is not obscure. It has been in the public domain since at least Wald's wartime work, formally named and categorised in the academic literature since at least the early 1990s, and popularised in accessible form by books from Taleb's The Black Swan (2007) to Ellenberg's How Not to Be Wrong (2014). The Abraham Wald story has been told in thousands of contexts: business school classrooms, TED talks, rationalist forums, popular science podcasts, management training sessions. It is, in the vocabulary of the internet, extremely well-known.
And yet the behaviour it describes remains almost entirely unchanged.
Why?
The first reason is structural. The evidence that would correct survivorship bias is, by definition, invisible. To see the missing planes, you would have to have access to data that does not exist — the performance records of funds that closed before they entered any database, the career trajectories of people who gave up before attracting notice, the clinical outcomes of patients who dropped out before a trial concluded. In a few academic contexts, researchers have gone to the effort of reconstructing this missing data, and the results are consistently sobering: adjusting for survivorship bias almost always reduces apparent performance, apparent efficacy, and apparent causation. But this reconstruction requires deliberate institutional effort, significant resources, and a motivation to find uncomfortable answers — a motivation that the people best positioned to undertake the work often lack.
The second reason is incentive. The financial industry has a commercial interest in showing performance data on a basis that flatters the survivor. The business book has a commercial interest in studying companies that succeeded rather than companies that failed, because success is aspirational and failure is merely instructive, and aspirational books outsell instructive ones. The commencement speaker has a social interest in offering encouragement rather than base rates. The mentor has a psychological interest in presenting their own path as a model rather than a single data point in a large and noisy distribution. These are not cynical interests, necessarily. They are structural features of a situation in which the production and dissemination of advice is shaped by incentives that point towards the visible and the positive.
The third reason is cognitive, and it is the most durable. The availability heuristic is not a bug in human cognition that can be patched by exposure to the concept. It is a feature — a computational efficiency that evolved because, in most ancestral environments, visible evidence was a reliable guide to reality. The tiger you could not see was not, usually, a tiger you needed to worry about. The food that looked plentiful actually was. The social dynamics that felt stable generally were. In a world of small groups, local interactions, and direct feedback, the systematic distortions introduced by selection effects operating on large populations of invisible data points simply did not arise at evolutionary scale.
What Wald understood — and what the cognitive science of the past fifty years has confirmed — is that modern reasoning requires a deliberate compensatory mechanism for the heuristics that served us well in other environments. The mechanism is not automatic. It has to be installed and maintained by something very like an active practice of statistical imagination: the discipline of asking, regularly and systematically, what is not here that should be.
VIII. A Practical Guide to Counting the Missing
Wald's genius was not, finally, a matter of knowing the concept. He did not merely know about survivorship bias in the sense of having been told about it or having read a paper. He had internalised the logical structure of the problem deeply enough that when he was presented with data, his first instinct was to ask what the data excluded. This is a different kind of knowing — what the philosopher Gilbert Ryle called "knowing how" as opposed to "knowing that." It is the difference between knowing that bicycles require balance and knowing how to balance on a bicycle.
What follows is not a list of heuristics — heuristics are themselves subject to survivorship bias, as the ones that get published and repeated are those that worked in the visible cases. It is, rather, a set of questions to be asked as a matter of disciplined routine.
Ask: Who is not in this room?
Before taking advice from any source, identify the population from which that source was drawn and ask who from that population is absent. If you are taking career advice from a successful professional, ask what proportion of people who followed the same path arrived at a different outcome, and ask whether those people were invited to speak. If you are consulting investment performance data, ask how many funds existed at the beginning of the measurement period and how many are represented in the final data. If you are reading a business book, ask the author to name and seriously analyse five companies that followed the same principles described in the book and failed.
Seek the graveyard.
This is Taleb's prescription: go deliberately looking for the cases that did not survive. In finance, this means using databases specifically constructed to include defunct funds and de-listed securities. In business, it means studying failure cases with the same rigour applied to success cases — not as cautionary moral tales, but as data. In medicine, it means looking for pre-registered trials that were never published and registered studies with null results. In personal decision-making, it means actively seeking out people who pursued the path you are considering and did not arrive where you hope to go.
Invert the question.
Instead of asking "What do the successful have in common?", ask "What do the failures have in common?" This is not merely a rhetorical reversal. It frequently produces different and more useful answers. The traits that characterise survivors often characterise failures in equal proportion. Risk-taking, for instance, is shared by almost all startups, both successful and unsuccessful. Passion is present in many careers, both rewarding and unrewarding. The distinguishing factors may lie not in the shared positive traits but in the structural conditions — timing, network, capital, market readiness, luck — that the success narrative tends to minimise.
Adjust for base rates.
The base rate is the prior probability of success in a given domain before any individual-level information is considered. In venture capital, the base rate for individual startups achieving meaningful returns is somewhere between 1 and 10 per cent, depending on the study and the definition of success. In creative careers, the base rate for achieving financial sustainability through art, music, or writing alone is similarly low. None of this means those careers should not be pursued — but decisions made without reference to the base rate are decisions made in a fantasy world populated only by survivors.
Embrace the silence as data.
The absence of evidence is frequently evidence of absence — or at least evidence of a selection effect. If a strategy, a treatment, a business model, or a career path is not represented in the successful cases, that absence is informative. It is not definitive — the absence might reflect a failure of imagination or opportunity rather than of the strategy itself — but it is a signal that demands explanation rather than dismissal.
IX. The Discipline of the Missing
There is a final dimension to Wald's insight that tends to get lost in the retelling. The story is usually told as a demonstration of statistical cleverness — the brilliant analyst who saw what the generals missed. But there is something else in it, something less comfortable.
The generals were not stupid. They were educated, experienced, professionally competent men who had spent their careers thinking rigorously about military problems. They had access to exactly the same data as Wald. They had the motivation — the highest possible motivation, measured in the lives of bomber crews — to reason correctly. And they got it wrong. Not because they were careless, but because the cognitive architecture of human reasoning, when applied to problems involving selection effects and invisible data, produces exactly the kind of error they made.
This should give us pause. If professional military analysts under wartime conditions, with lives at stake, are not immune to survivorship bias, then neither are we. The assumption that knowing the concept protects us against the error is itself a form of optimism that Wald's story should make us sceptical of — a survivor bias about the effectiveness of bias education.
What Wald had, beyond knowledge of the mathematical structure of the problem, was institutional authority to ask the uncomfortable question and to insist on an answer. He was in the room. He was listened to. He had the methodological standing to redirect the analysis. In most real-world contexts, the person who points out that the data set is biassed, that the evidence on the table represents only the survivors, that the missing cases deserve to be counted, is asking the organisation to do more work in order to arrive at a less flattering conclusion. This is rarely welcomed.
The practical implication is that correcting for survivorship bias in institutional contexts requires not just statistical literacy but organisational courage: the willingness to commission and publish research that will find what the comfortable version of the data does not show, to invite the failures to speak alongside the successes, to design selection processes that document and learn from the cases that did not make it through. This is expensive, uncomfortable, and structurally unrewarded. It is also the work.
X. Conclusion: The Plane That Is Not on the Chart
Eight decades have passed since Abraham Wald sat down with a chart of bullet holes and asked where the planes were that should have been on it. In those eight decades, the academic literature on survivorship bias has grown to encompass finance, medicine, economics, psychology, epidemiology, ecology, and management science. The mathematical structure of the problem has been precisely characterised. The cognitive mechanisms that make the error feel natural — the availability heuristic, the narrative fallacy, Kahneman's WYSIATI (What You See Is All There Is) — have been mapped in granular detail.
And the error persists. It persists in the investment brochures that show you only the funds that survived. It persists in the management books that study only the companies that succeeded. It persists in the career advice dispensed by the people who made it to career advisers. It persists in the clinical trials that get published because they found an effect, and in the ones that do not get published because they did not. It persists in the self-help industry, which is itself a spectacular demonstration of survivorship bias — an industry built on the testimony of people for whom a particular kind of self-help happened to work, selling books and courses to people who want to replicate outcomes from a sample of one.
The error persists, ultimately, because it is built into the structure of visibility itself. We see what is present. We infer the absent. And in a world where the absent are systematically different from the present — systematically less successful, less visible, less commercially tractable — our inferences will be systematically wrong in the same direction every time.
Wald did not need a computer. He did not need an algorithm, a data warehouse, or a machine learning pipeline. He needed a pencil, a clear head, and the discipline to ask a simple question: what is not here that, if the world were as I am being asked to assume, should be?
That question is free to ask. It has been free for eighty years. Most people do not ask it, not because they cannot, but because the answer is uncomfortable, expensive to obtain, and structurally unrewarded in most of the environments we inhabit.
Next time you are handed a list of winners — whether it concerns a stock, a strategy, a career path, a business model, or an approach to living — pause before you study the bullet holes. Look first for the planes that did not come back. Their absence is the most important data in the room.
Bibliography and Further Reading
Primary Sources and Academic Studies
- Wald, Abraham. A Method of Estimating Plane Vulnerability Based on Damage of Survivors. Statistical Research Group, Columbia University, 1943; reprinted by Center for Naval Analyses, 1981.
- Brown, Stephen J., William N. Goetzmann, Roger G. Ibbotson, and Stephen A. Ross. "Survivorship Bias in Performance Studies." Review of Financial Studies 5, no. 4 (1992): 553–580.
- Brown, Stephen J., and William N. Goetzmann. "Performance Persistence." Journal of Finance 50, no. 2 (1995): 679–698.
- Carhart, Mark M. "On Persistence in Mutual Fund Performance." Journal of Finance 52, no. 1 (1997): 57–82.
- Elton, Edwin J., Martin J. Gruber, and Christopher R. Blake. "Survivorship Bias and Mutual Fund Performance." Review of Financial Studies 9, no. 4 (1996): 1097–1120.
- Malkiel, Burton G. "Returns from Investing in Equity Mutual Funds 1971 to 1991." Journal of Finance 50, no. 2 (1995): 549–572.
- Mangel, Marc, and Francisco J. Samaniego. "Abraham Wald's Work on Aircraft Survivability." Journal of the American Statistical Association 79, no. 386 (1984): 259–267.
- Tversky, Amos, and Daniel Kahneman. "Availability: A Heuristic for Judging Frequency and Probability." Cognitive Psychology 5, no. 2 (1973): 207–232.
- Shukla, Rishi, et al. "Survivorship Bias in Emerging Market Small-Cap Indices: Evidence from India's NIFTY Smallcap 250." arXiv preprint arXiv:2603.19380 (2025).
Books
- Collins, Jim, and Jerry I. Porras. Built to Last: Successful Habits of Visionary Companies. Harper Business, 1994.
- Collins, Jim. Good to Great: Why Some Companies Make the Leap and Others Don't. Random House Business, 2001.
- Ellenberg, Jordan. How Not to Be Wrong: The Power of Mathematical Thinking. Allen Lane, 2014.
- Kahneman, Daniel. Thinking, Fast and Slow. Farrar, Straus and Giroux, 2011.
- Malkiel, Burton G. A Random Walk Down Wall Street. W. W. Norton, 1973 (revised multiple editions).
- Mlodinow, Leonard. The Drunkard's Walk: How Randomness Rules Our Lives. Pantheon Books, 2008.
- Newport, Cal. So Good They Cannot Ignore You: Why Skills Trump Passion in the Quest for Work You Love. Business Plus, 2012.
- Rosenzweig, Phil. The Halo Effect: How Managers Let Themselves Be Deceived. Simon & Schuster, 2007.
- Schwartz, Barry. The Paradox of Choice: Why More is Less. Ecco, 2004.
- Silver, Nate. The Signal and the Noise: Why So Many Predictions Fail — but Some Don't. Penguin Press, 2012.
- Taleb, Nassim Nicholas. Fooled by Randomness: The Hidden Role of Chance in Life and in the Markets. Texere, 2001.
- Taleb, Nassim Nicholas. The Black Swan: The Impact of the Highly Improbable. Random House, 2007.
- Taleb, Nassim Nicholas. Antifragile: Things That Gain from Disorder. Random House, 2012.
Further Academic Context
- Ericsson, K. Anders, Ralf Th. Krampe, and Clemens Tesch-Römer. "The Role of Deliberate Practice in the Acquisition of Expert Performance." Psychological Review 100, no. 3 (1993): 363–406.
- Macnamara, Brooke N., David Z. Hambrick, and Frederick L. Oswald. "Deliberate Practice and Performance in Music, Games, Sports, Education, and Professions: A Meta-Analysis." Psychological Science 25, no. 8 (2014): 1608–1618.
- Joannidis, John P. A. "Why Most Published Research Findings Are False." PLOS Medicine 2, no. 8 (2005): e124.
Petru Cojocaru writes on personal development, critical thinking, and digital culture at cojocarupetru.info. He is the founder of ABSOLUT WEB EXPERT SRL, a web agency based in Botoșani, Romania.