The words we use to express probability are far more slippery than we realise. The consequences, from the trading floor to the boardroom, can be enormous

On the morning of March 24, 2026, Donald Trump stood before cameras in Tennessee and told reporters there was “a very good chance” the United States and Iran would reach a deal to end a war that had already killed more than 2,000 people. Brent Crude, which had been trading above $118 a barrel at the peak of the conflict, fell by more than 10% that same day.

Global equity markets surged. Bitcoin jumped 3%. Bonds rallied. All of this from four words: a very good chance.

What, exactly, is a very good chance? Is it 60%? 75%? 90%? The answer is that nobody really agrees. And that shared misunderstanding, multiplied across millions of traders, analysts, diplomats and ordinary citizens, has consequences that range from the financially catastrophic to the culturally confusing.

Everybody interprets probability words differently

In 2020, Sanne Willems, Casper Albers, and Ionica Smeets at Leiden and Groningen universities published a study in the Journal of Science Communication on how people interpret probability phrases. They surveyed 881 Dutch speakers, both statisticians and laypeople, on how they numerically interpreted common probability phrases. The results were striking.

When asked what percentage they assigned to the word ‘probable’, respondents gave answers ranging from 41% to 86%. Even ‘almost always’, which sounds about as clear as language gets, stretched from 70% to 96% in the minds of different readers.

Perhaps most surprising of all, statisticians who spend their professional lives working with probability, showed almost no more consensus than everyone else. Professional expertise, it turns out, does not inoculate you against the ambiguity of ordinary language.

The one area of relative agreement? Extremes. Words like certain, always, never and impossible produced narrower bands of interpretation. The moment language enters themiddle ground, the contested territory where most real decisions are made, it becomes genuinely unreliable as a signal.

The Trump Variable

No one has demonstrated this more vividly, or more expensively, in recent months than the 47th President of the United States. Trump’s probabilistic vocabulary is in a class of its own: “very good chance,” “serious chance,” “good chance,” “maybe,” “probably,” “we’ll see,” “could very well happen”, a rotating lexicon of deliberate vagueness that he has deployed across Iran, tariffs, Ukraine and NATO with breathtaking consistency.

The financial consequences have been measurable and enormous. When Trump told reporters in late March that he had “a very serious chance of making a deal” with Iran and pointedly added, “that doesn’t guarantee anything, I’m not guaranteeing anything,” markets had already moved billions of dollars on the basis of the words that came before the caveat. His own acknowledgement of the problem, “I’m not going to come out here in a week or two weeks and have you all say, ‘Oh, you said…’” was itself a masterclass in how probability language can be weaponised as plausible deniability.

The European Food Safety Authority has a formal table that attempts to map verbal probability expressions to numerical ranges. Very likely means above 95%, likely means 66-95%, and so on, precisely because organisations understand the danger of the gap.

No such table governs the Oval Office. When Trump says “I think there’s a very good chance,” he is communicating simultaneously to markets, allies, adversaries and a domestic audience, each of whom will map the words onto different numbers and act accordingly.

The ambiguity is deliberate.

Don’t listen to what I said, listen to what I meant

If probability language means different things within a single language, crossing cultural and linguistic lines amplifies the range of interpretations. The Leiden study noted that earlier research comparing English, French, German and Chinese probability expressions found that even careful translations shifted numerical interpretations significantly. French and Chinese speakers showed more variability than English speakers, not less.

This is where INSEAD professor Erin Meyer’s work becomes essential reading for anyone managing international teams or cross-border negotiations. Meyer, whose Culture Map framework has become a widely used tools in global executive education, distinguishes between what she calls low-context and high-context cultures.

In low-context cultures – the United States an extreme example in her research – communication is expected to be explicit, direct, and self-contained. Words mean what they say.

In high-context cultures – Japan, Korea, much of the Arab world, and significant parts of Europe – meaning lives between the lines. Shared context, relationships, and unspoken subtext carry at least as much weight as the literal content of what is said.

The implications for probability language are profound. A Japanese executive who says a deal is “possible” is likely encoding a polite but firm reluctance. An American saying the same word probably means it could happen, 50-50. A French executive using “il est probable que” may be expressing something closer to confident expectation, while a Dutch counterpart using the closest equivalent may be signalling genuine uncertainty.

Meyer describes a French concept, sous-entendu, that captures this beautifully: “don’t listen to what I said, listen to what I meant.” In high-context cultures, probability language is never just probability language. It is a social signal, a relationship marker, and a face-saving device all at once.

The Boardroom Blind Spot

These dynamics play out constantly in business, often without anyone noticing. Consider a common scenario: a multinational team is assessing whether a product launch in a new market is viable. The American country manager says it is “likely to succeed.” The Korean finance director says success is “not impossible.” The German operations lead says they “cannot rule out significant challenges.” 

Each person believes they have communicated a meaningful probability. The American walks away thinking the team is broadly aligned. The Korean and German leads walk away believing they have registered serious reservations. The meeting ends. The product launches. The failure is later attributed to poor execution rather than catastrophic miscommunication at the planning stage.

Meyer’s advice for leaders navigating this terrain is practical and concrete: in mixed cultural teams, make it routine to attach numbers to probability language. Not as a constraint on discussion, but as a clarification. If someone says “likely,” ask, “When you say likely, are we talking 60% or 85%?” 

The question is not pedantic. It is the difference between shared understanding and comfortable illusion.

There is also, she notes, a directional asymmetry in how cultures handle uncertainty itself. Cultures that score high on what the Dutch social psychologist, Geert Hofstede called “uncertainty avoidance” – Finland, Germany, Japan, South Korea – have a strong preference for clear frameworks, written rules and structured communication precisely because ambiguity is uncomfortable.

Cultures lower on that scale such as Singapore, Sweden, the UK are more at ease with open-ended situations and looser language. Put a German engineer and a British entrepreneur in the same room to assess project risk and you have not just a vocabulary problem but a philosophical one.

Medicine, Climate, and the Stakes We Ignore

The stakes are not confined to financial markets or international diplomacy. The Willems study was motivated in part by a disturbing finding in health communication: when a drug’s side effect was described to patients as common – a term that medical guidelines define as affecting 1-10% of users – patients estimated on average that 34% of people would experience it. 

The result was measurable harm: reduced medication adherence, and real people experiencing side effects simply because they had been told to expect them.

Climate science has faced the same problem. The Intergovernmental Panel on Climate Change has its own probability vocabulary – likely, very likely, virtually certain – mapped to specific percentage ranges. Research across multiple countries has shown that readers in China and France interpret these terms quite differently from theintended meanings, systematically overestimating or underestimating the probability of warming scenarios depending on which linguistic conventions they bring to the text.

The Leiden researchers’ recommendation is to always pair verbal probability expressions with numerical values. “A very good chance (approximately 75-85%)” is more honest than “a very good chance” alone, even if it is less rhetorically graceful.

What Clear Communication Looks Like

None of this means probabilistic language should be abandoned. It has genuine advantages. As the Leiden study notes, verbal phrases acknowledge uncertainty in ways that a confident-sounding percentage often conceals. A doctor who says there is, “a reasonable chance this treatment will work” is communicating something subtly different from one who says, “there is a 65% chance this treatment will work,” even if the underlying probability is identical.

The first acknowledges the limits of prediction. The second can invite false precision.

The goal, then, is not to replace language with numbers, but to develop the self-awareness to know when the gap between your intended meaning and your listener’s interpretation is likely to be large, and to close it deliberately.

For business leaders, Meyer offers a practical framework. Before any important communication involving probability or risk, ask yourself two questions. First, is my listener coming from a culture with significantly different communication norms? Second, is the decision they are about to make sensitive to the difference between, say, 60% and 80%? If the answer to both is yes, slow down and be explicit.

For negotiators – and here Trump’s Iran communications are instructive – probabilistic vagueness is sometimes a tool. “There is a very good chance of a deal” preserves optionality, signals goodwill, and moves markets, all without any commitment. The danger is that the other side in a negotiation, especially one from a high-context culture, may interpret the same words as a stronger signal than they were meant to be, and feel deceived when the deal doesn’t come.

Reading the Air

Willems and her colleagues recommend that science communicators, and by extension, anyone making important decisions on behalf of others should treat verbal probability expressions with considerable caution. The same care applies in the executive suite, on the trading floor, and in the Cabinet room.

When the words “very good chance” move oil markets by 10%, the gap between what a speaker means and what listeners hear is not an abstract linguistic curiosity. It is a market risk, a diplomatic liability, and sometimes a matter of life and death. The lesson is not that language is broken, but that it is shaped by culture, context, and the particular cognitive shortcuts of every individual who encounters it.

The next time someone tells you that something is “likely,” it might be worth asking what they actually mean. You may be surprised, or alarmed by the answer.

About the author

Matt Symonds is Chief Editor of BlueSky Thinking, and host of BlueSky Media Connect, bringing together b-schools and universities to meet editors from FT, BBC, Bloomberg, WSJ, The Economist, NYTimes and other global / regional media.

He is the S of QS, co-founding QS Quacquarelli Symonds, publishers of the QS World University Rankings. Matt I also co-Founder and Director of Fortuna Admissions, a coaching dream team of former business school and university admissions professionals from top-tier institutions, including Harvard, Stanford, Wharton, INSEAD, LBS, Chicago Booth, Columbia, Northwestern Kellogg, Berkeley Haas.

Matt co-host the CentreCourt MBA & Masters Festivals with John A. Byrne and Poets & Quants. Author of the international bestseller, “Getting the MBA Admissions Edge” sponsored by Goldman Sachs, McKinsey, Bain, BCG, he writes about Higher Education and management for BBC, Times of India and formerly Forbes, The Economist and Bloomberg.

Interested in this topic? You might also like…