Loading module…
Loading module…
Module 1 of 6
Every decision that matters is a decision made without knowing how it will turn out. If you knew the outcome, there would be no decision to make — you would simply take the action that leads to the outcome you want. Decision-making is the act of choosing under uncertainty, and the quality of your decisions is largely determined by how well you understand and handle the not-knowing at their core. This course is about that skill: making good decisions when you cannot know how they will turn out, which is to say, making good decisions at all.
The foundation, and the subject of this first module, is that uncertainty is not one thing. The word covers several fundamentally different situations that call for fundamentally different responses, and one of the most common and expensive errors in decision-making is treating one type of uncertainty as if it were another — applying the tools for measurable risk to a situation of deep unknowns, or freezing in the face of a manageable risk as if it were unfathomable, or mistaking a situation you could resolve with information for one you cannot. Before you can handle uncertainty well, you must be able to tell what kind of uncertainty you are facing, because the kind determines the response. This module builds that discrimination.
By the end you will be able to distinguish risk (measurable, quantifiable uncertainty) from true uncertainty (unmeasurable, unquantifiable); recognize the difference between aleatory uncertainty (irreducible randomness) and epistemic uncertainty (reducible ignorance) and why it dictates whether to gather information or accept the odds; understand the known/unknown matrix and the special danger of unknown unknowns; grasp why the human mind systematically mishandles uncertainty by converting it into false certainty; and match the type of uncertainty to the appropriate decision approach. This is the diagnostic layer of decision-making — figuring out what you are actually dealing with — and it precedes and governs every technique in the rest of the course.
The most fundamental distinction in the study of uncertainty was drawn by the economist Frank Knight in 1921, and it separates two things the everyday word "uncertainty" runs together: risk and true uncertainty. The distinction is not academic; it determines which entire toolkit applies, and confusing the two is the source of some of the most consequential decision errors, including some that have brought down companies and crashed markets.
Risk is uncertainty that can be measured — situations where you do not know the outcome, but you do know the possible outcomes and can assign probabilities to them, whether from theory, from data, or from repeated experience. The roll of a fair die is risk: you don't know what will come up, but you know it will be one through six, each with probability one-sixth. An insurance company's payout on a large pool of policies is risk: it cannot predict which individual will make a claim, but it knows from vast data the probability distribution of claims across the pool, precisely enough to price the policies and profit. Risk is the domain of probability, statistics, expected value, and the entire mathematical apparatus this course will develop in later modules — because those tools require known probabilities to work, and risk is exactly the situation where the probabilities are known or knowable.
True uncertainty (Knight called it simply "uncertainty," and it is sometimes called Knightian uncertainty or ambiguity) is uncertainty that cannot be measured — situations where you do not know the possible outcomes, or cannot assign meaningful probabilities to them, because the situation is unique, unprecedented, too complex, or genuinely unknowable. The success of a genuinely novel product in a market that does not yet exist is true uncertainty: there is no probability distribution to look up, no data from repeated trials, no theory that yields the odds, because the situation is one of a kind. The behavior of a complex system undergoing an unprecedented shock is true uncertainty. Most of the important decisions in life and strategy — whom to marry, what career to pursue, whether to launch a novel venture, how a war or a technology or a society will unfold — are closer to true uncertainty than to risk, because they are unique, unrepeatable situations for which no reliable probability distribution exists.
The critical error is treating true uncertainty as if it were risk — manufacturing precise-looking probabilities for situations where no real probabilities exist, and then trusting the mathematical apparatus built on those false numbers. This error is seductive because the tools of risk are powerful and satisfying, and applying them produces confident-looking outputs — expected values, confidence intervals, risk models — that feel like knowledge. But when the underlying probabilities were fabricated for a situation of true uncertainty, the confident outputs are false precision: numbers that look like measurement but are actually guesses dressed in mathematics, and trusting them is more dangerous than admitting you don't know, because the false precision hides the ignorance it should reveal. Much of the 2008 financial crisis has been attributed to exactly this: risk models that assigned precise probabilities to mortgage defaults, treating as measurable risk what was actually true uncertainty about an unprecedented housing market, so that the confident numbers masked the real unknowability until it manifested catastrophically.
The discipline is to know which you are facing before choosing your tools. Ask: do I actually know the possible outcomes and their probabilities, from real data, sound theory, or genuine repeated experience? If yes, it is risk, and the probabilistic tools apply. If no — if I am estimating probabilities for a unique, unprecedented, or unmeasurable situation — it is true uncertainty, and the probabilistic tools apply only as rough, clearly-labeled aids, never as measurements, and other approaches (robustness, optionality, humility, reversibility) must carry the weight. The rest of this course develops the tools of risk in full, but this module plants the flag that must never be forgotten: those tools require real probabilities, and applying them to true uncertainty produces dangerous false precision. Knowing the difference is the first competence.
Key Insight: Knight's distinction separates risk (uncertainty with knowable outcomes and probabilities, from data, theory, or repeated experience — the domain of the probabilistic tools) from true uncertainty (unique, unprecedented, or unmeasurable situations where no real probabilities exist). The critical error is treating true uncertainty as risk — manufacturing precise probabilities that produce confident-looking outputs masking genuine ignorance, more dangerous than admitting you don't know. Most important decisions are closer to true uncertainty, so the first competence is knowing which you face before choosing your tools.
A second distinction cuts across the first and answers a different, equally practical question: can this uncertainty be reduced by gathering information, or not? The answer determines whether the right response is to go get more information before deciding, or to stop seeking information and act on the odds. This is the distinction between aleatory and epistemic uncertainty, and getting it wrong wastes resources seeking information that cannot help, or gambles on odds when information could have resolved them.
Aleatory uncertainty (from the Latin for "dice") is uncertainty arising from inherent randomness in the situation — variability that is irreducible, that no amount of additional information can eliminate, because it is a feature of the process itself rather than of your knowledge. The outcome of a future fair coin flip is aleatory: no amount of study can tell you how it will land, because the uncertainty is in the random process, not in your ignorance. The specific value a genuinely random process will produce next is aleatory. Aleatory uncertainty can be characterized (you can know the coin is fair, so the probability is one-half) but not reduced — you can know the odds precisely and still not know the outcome, because the outcome is inherently random. When uncertainty is purely aleatory, gathering more information about the specific outcome is futile; the correct response is to accept the irreducible randomness, characterize it as well as possible (get the odds right), and make the decision on the basis of the odds, because there is nothing more to learn about the specific result.
Epistemic uncertainty (from the Greek for "knowledge") is uncertainty arising from your lack of knowledge — from ignorance that could, in principle, be reduced or eliminated by gathering more information, because the fact is determinate and merely unknown to you. Whether a particular already-flipped coin, now covered, landed heads or tails is epistemic: the outcome is fixed and knowable, you just don't know it, and lifting the cover (gathering information) resolves the uncertainty completely. The correct diagnosis for a patient, the true state of a competitor's finances, the actual cause of a malfunction — these are epistemic uncertainties: there is a fact of the matter, and investigation can reveal it. When uncertainty is epistemic and the information is worth its cost, the correct response is to reduce the uncertainty by gathering information before deciding, because unlike aleatory uncertainty, this uncertainty can be resolved, and deciding in ignorance when you could cheaply have known is a real error.
The practical power of the distinction is that it tells you where to spend effort. Facing an uncertainty, ask: is this irreducible randomness (aleatory) or resolvable ignorance (epistemic)? If aleatory, do not waste resources seeking information about the specific outcome — it cannot be had; instead characterize the odds and decide on them. If epistemic, and the information is affordable and would change the decision, get the information — the uncertainty is reducible and deciding without reducing it is leaving value on the table. The error in one direction is treating epistemic uncertainty as aleatory — gambling on the odds when you could have simply found out, deciding in ignorance you could have cheaply cured. The error in the other direction is treating aleatory uncertainty as epistemic — endlessly seeking information to resolve what is inherently random, delaying the decision in pursuit of a certainty that does not exist to be found, the analysis-paralysis of trying to learn away irreducible randomness.
Most real uncertainties are mixtures of both, and the sophistication is to separate the components: which part of this uncertainty is reducible ignorance I should invest in resolving, and which part is irreducible randomness I must simply accept and decide upon? The uncertainty about whether a new product will succeed contains epistemic components (what do customers actually want? — resolvable by research) and aleatory components (how will an unpredictable market react? — irreducibly random). The skilled decider resolves the epistemic part up to the point where information stops being worth its cost, then accepts the aleatory remainder and decides on the odds — rather than either gambling on the whole thing without doing the resolvable research, or endlessly researching in pursuit of resolving the irreducibly random part. This is the discipline of knowing what more information can and cannot buy you, and it is one of the most practically valuable distinctions in the whole subject.
Key Insight: Aleatory uncertainty is irreducible randomness inherent in the process (no information reduces it — characterize the odds and decide on them); epistemic uncertainty is resolvable ignorance about a determinate fact (gather information to resolve it, if worth the cost). The distinction tells you where to spend effort: don't seek information about the irreducibly random (analysis paralysis), and don't gamble on odds you could cheaply resolve by finding out (deciding in curable ignorance). Most real uncertainties mix both — resolve the epistemic part up to the point information stops being worth its cost, then accept and decide on the aleatory remainder.
A third framework, popularized in decision and risk circles and famously articulated in a 2002 briefing, organizes uncertainty by a different axis: not whether it is measurable or reducible, but whether you are aware of it at all. This is the known/unknown matrix, and its most important cell — the unknown unknowns — is the source of the uncertainties that do the most damage precisely because they are not on anyone's radar.
The matrix crosses two questions: do you know about a thing, and do you know the thing? This yields four quadrants. Known knowns are things you know that you know — the facts you are aware of and have correct: the established, understood elements of the situation. These pose no uncertainty; they are your solid ground. Known unknowns are things you know that you don't know — the uncertainties you are aware of: you know there is a question, you know you lack the answer, and you can therefore investigate it, plan for it, or explicitly account for it. "We don't know how the market will react" is a known unknown — an uncertainty you can see and address. Known unknowns are manageable precisely because you know they are there; they are the uncertainties decision-making normally handles.
Unknown unknowns are things you don't know that you don't know — the uncertainties you are not even aware exist, the possibilities that never entered your consideration, the factors you did not know were factors. These are the most dangerous class of uncertainty by far, because you cannot investigate, plan for, or account for a risk you do not know exists — it is entirely outside your model of the situation, so it is not in your analysis, not in your contingency plans, not in your risk assessment, until it manifests and blindsides you. The catastrophic surprises — the failure mode no one imagined, the risk that was not on any list, the event that "no one saw coming" — are unknown unknowns. They are terrifying precisely because their defining feature is invisibility in advance: by definition, you do not know to look for them. (The fourth quadrant, unknown knowns — things you don't know that you do know, tacit knowledge or things you've suppressed or forgotten — is subtler and less central, but it captures the phenomenon of an organization "knowing" something collectively that no decision-maker has surfaced.)
The strategic importance of the matrix is what it tells you to do about each class. Known knowns: rely on them, but verify they are actually true (Section 5's warning about false certainty — many supposed known knowns "just ain't so"). Known unknowns: investigate them where epistemic and worthwhile, plan contingencies for them, and account for them explicitly in the decision. Unknown unknowns: here the situation is different, because you cannot address a specific unknown unknown — you don't know what it is. What you can do is (a) work to convert unknown unknowns into known unknowns by actively searching for the risks you haven't thought of — asking "what am I not considering? what would surprise me? what is outside my model?" — which is exactly the function of techniques like the pre-mortem (a whole later module), red-teaming, and seeking outside perspectives that see what your frame excludes; and (b) build general robustness against the unforeseen — margins of safety, reserves, reversibility, avoiding fragility — so that when an unknown unknown does manifest (as some always will, because you can never surface them all), you survive it. You cannot predict the specific unknown unknown, but you can decide in a way that does not get destroyed by the mere existence of unforeseen possibilities.
The deepest lesson of the matrix is humility about the completeness of your own model. The most dangerous decision-maker is the one who believes they have accounted for everything — because that belief is precisely what blinds them to the unknown unknowns, which by definition lie outside the accounting. The decision-maker who knows their model is incomplete, who actively hunts for the risks they haven't thought of, and who builds robustness against the surprises they cannot foresee, is far safer than the one whose confidence has convinced them there is nothing left to find. In uncertainty, the awareness of your own blindness is a form of sight.
Key Insight: The known/unknown matrix classifies uncertainty by awareness: known knowns (solid ground, but verify), known unknowns (visible uncertainties you can investigate and plan for), and — most dangerous — unknown unknowns (risks outside your model entirely, which blindside you precisely because you don't know to look for them). You cannot address a specific unknown unknown, but you can hunt to convert them into known unknowns (pre-mortems, red-teaming, outside views) and build general robustness (margins, reserves, reversibility) to survive the ones you can't foresee. The most dangerous decider is the one who believes they've accounted for everything — that belief is what blinds them.
Beyond the formal categories, uncertainty comes in distinguishable textures that call for different responses, and a useful vocabulary for these has emerged in the form of the VUCA framework (Volatility, Uncertainty, Complexity, Ambiguity) and related distinctions. Setting aside the acronym's looseness, three textures are worth distinguishing precisely, because each is a different reason you can't predict the outcome and each has a different remedy.
Volatility is uncertainty from rapid, large, and unpredictable change in a situation whose basic structure you do understand. The situation is not mysterious — you know how it works — but it is changing fast and swinging widely, so that even with good understanding you cannot predict the specific state at a given future moment because it moves too fast and too far. A market you understand well but that is swinging violently is volatile. The remedy for volatility is not more information about the underlying structure (you already understand it) but agility and slack: the ability to respond quickly to changes as they come, and reserves and buffers that let you absorb the swings without being forced into bad decisions at bad moments. You handle volatility by being responsive and cushioned, not by trying to predict the unpredictable swings.
Complexity is uncertainty from many interacting parts whose interactions produce behavior that is hard to predict even when each part is understood, because the whole behaves in ways not readily deducible from the parts (emergence), with feedback loops, nonlinearities, and cascades. The situation is not random and not necessarily fast-changing, but it is tangled: so many interconnected elements that tracing cause to effect is beyond your (or anyone's) analytical reach. A large ecosystem, a global supply chain, an economy, a complex organization are complex. The remedy for complexity is not to pretend you can fully model it (you can't) but to simplify where possible, experiment to learn how the system responds, decentralize decisions to those closest to each part, and watch for the leverage points and feedback loops rather than trying to compute the whole. You handle complexity by probing and responding rather than by predicting, because the system is too tangled to predict but often can be learned about through interaction.
Ambiguity is uncertainty from not understanding the basic nature of the situation — where you don't even know the rules, the cause-and-effect relationships, or how things work, because the situation is genuinely novel or the information is fundamentally unclear. This is the deepest texture, close to Knight's true uncertainty: not that you can't predict a well-understood system's fast swings (volatility) or trace a tangled system's interactions (complexity), but that you don't grasp what kind of situation you're in at all. A genuinely unprecedented situation, a technology whose implications no one understands yet, is ambiguous. The remedy for ambiguity is experimentation and hypothesis-testing to build understanding — you cannot analyze your way to clarity because you lack the basic model, so you must act (carefully, reversibly) to generate the information that builds the model, learning what kind of situation you're in by engaging it. You handle ambiguity by experimenting to understand, treating early moves as probes that reveal the nature of the situation rather than as bets on a situation you understand.
The practical value of distinguishing these textures is that each feels like "I can't predict this" but each has a different cause and cure. Diagnosing the texture tells you the remedy: volatility calls for agility and slack, complexity for probing and decentralizing, ambiguity for experimenting to build understanding. Misdiagnosing wastes the remedy — trying to build more understanding of a volatile situation you already understand (the problem is speed, not comprehension), or trying to respond with agility to an ambiguous situation whose nature you don't yet grasp (you'll respond fast in the wrong direction). Reading the texture correctly is part of the diagnostic competence this module builds, and it connects the abstract categories to the concrete question every decider faces: why can't I predict this, and therefore what should I do about it?
Key Insight: Uncertainty comes in distinguishable textures with different cures: volatility (fast, wide swings in a structure you understand — remedy: agility and slack, not more information), complexity (many tangled interacting parts producing hard-to-predict emergent behavior — remedy: probe, experiment, decentralize, watch leverage points), and ambiguity (you don't understand the situation's basic nature — remedy: careful reversible experimentation to build the model you lack). Each feels like "I can't predict this" but has a different cause, so diagnosing the texture prescribes the remedy — and misdiagnosing wastes it.
Understanding the types of uncertainty is necessary but not sufficient, because the human mind has powerful, systematic tendencies to mishandle uncertainty — chiefly by converting it into false certainty, because uncertainty is psychologically uncomfortable and certainty, even false certainty, feels better. Recognizing these tendencies in yourself is essential, because they operate automatically and against good decision-making, and the types of uncertainty are useless if the mind refuses to sit with uncertainty long enough to apply them.
The root of the problem is that uncertainty is aversive. The state of not-knowing, of holding open possibilities without resolution, is uncomfortable — it generates anxiety, it feels like weakness or incompetence, it resists the mind's drive to understand and control. This discomfort creates a powerful pull toward resolving the uncertainty prematurely, toward reaching a conclusion, any conclusion, that ends the uncomfortable open state — and a false certainty resolves the discomfort just as effectively as a true one, so the mind readily accepts false certainty as the price of comfort. This is the flight to false certainty, and it manifests in several specific cognitive tendencies.
Overconfidence: the systematic tendency to be more certain than the evidence warrants, to overestimate the accuracy of our knowledge and predictions. Studies of calibration consistently find people's "90% confident" judgments are right far less than 90% of the time — we are systematically overconfident, treating our estimates as more reliable than they are, which converts real uncertainty into felt certainty and leads us to decide as if we knew when we didn't. Premature closure: seizing on the first plausible answer and stopping the search, ending the uncomfortable uncertainty before adequately considering alternatives, so that the decision rests on an under-examined conclusion reached mainly to escape the discomfort of openness. Neglect of the unknown: the tendency to decide as if the known factors were the only factors, ignoring the unknown unknowns (Section 3) because they are invisible and uncomfortable to contemplate, building a confident decision on a model that omits everything outside it. Illusion of control: overestimating our ability to control or predict uncertain outcomes, treating aleatory (irreducible) uncertainty as if effort or skill could master it, so that we take on risks we falsely believe we can manage.
These tendencies interact with a social dimension: uncertainty is often punished and false certainty rewarded in human groups. The confident forecaster is believed and followed; the honest one who says "I don't know, here is the range of possibilities" is seen as weak or unhelpful. Leaders are expected to project certainty; expressing appropriate uncertainty can read as indecisiveness. This social reward for false certainty compounds the psychological one, pressuring even people who know better to convert their honest uncertainty into confident-sounding claims — which is how organizations end up acting on false certainties that everyone privately doubted. The Twain epigraph names the result: the trouble comes not from acknowledged ignorance but from the false certainties, the things "we know for sure that just ain't so," which we act on with full confidence precisely because we have converted our uncertainty into a comfortable, unwarranted conviction.
The corrective is a set of dispositions that run against the mind's grain and must be cultivated deliberately: tolerating the discomfort of uncertainty rather than fleeing it, holding open the not-knowing long enough to handle it well; calibrating confidence to evidence, deliberately reducing certainty toward what the evidence actually supports (later modules build the tools for this); actively seeking disconfirmation rather than premature closure, looking for what would prove your conclusion wrong before committing to it; and valuing honest uncertainty over false confidence, both in yourself and in those you listen to, recognizing that the person who says "I don't know" precisely is more trustworthy than the one who is confidently wrong. The whole rest of this course, in a sense, is tools for handling uncertainty well — but the tools only work if the mind will stay in the uncertain state long enough to use them, rather than fleeing immediately to the false certainty that feels better and decides worse. The first and most important corrective is the willingness to remain honestly uncertain.
Key Insight: The mind mishandles uncertainty by fleeing it — because not-knowing is aversive, it converts real uncertainty into comfortable false certainty through overconfidence (systematically more certain than evidence warrants), premature closure (seizing the first answer to end the discomfort), neglect of the unknown, and illusion of control — and groups compound this by rewarding confident forecasters and punishing honest "I don't knows." The trouble comes from the false certainties we act on with full confidence. The corrective runs against the grain: tolerate the discomfort, calibrate confidence to evidence, seek disconfirmation, and value honest uncertainty over false confidence.
Consolidating the module into a usable capability: you should now be able to diagnose what type of uncertainty a decision faces and therefore what approach it calls for — the diagnostic layer that governs every technique in the rest of the course. This closing section assembles the distinctions into a working diagnostic and traces it through an example, so that facing an uncertain decision, your first move is to correctly classify the uncertainty before reaching for any tool.
The diagnostic sequence. Facing an uncertain decision, run through the classifications: (1) Risk or true uncertainty? Do real probabilities exist (from data, theory, or repeated experience), or would any probabilities be fabricated for a unique situation? This determines whether the probabilistic tools apply as measurements or only as rough, clearly-labeled aids. (2) Aleatory or epistemic — and in what mixture? Which part of the uncertainty is irreducible randomness to be accepted and decided upon, and which part is resolvable ignorance worth gathering information to reduce? This determines where to spend effort seeking information. (3) Where does it sit on the known/unknown matrix? What are the known unknowns to investigate and plan for, and — critically — have you hunted for the unknown unknowns and built robustness against them? (4) What is the texture — volatility, complexity, or ambiguity? This determines whether the remedy is agility and slack, probing and decentralizing, or experimenting to build understanding. (5) Am I fleeing to false certainty? Check your own confidence against the evidence, and resist the pull to resolve the discomfort prematurely. The sequence produces a diagnosis: this kind of uncertainty, calling for this kind of approach.
Tracing the diagnostic through an example. A company is deciding whether to enter a new international market. Run the diagnostic. Risk or true uncertainty? Partly measurable (market size, demographics, existing-competitor performance — real data exists) and partly true uncertainty (how this product will fare in this novel context — unique, no real probability). So: use the probabilistic tools for the measurable parts, but do not fabricate precise probabilities for the novel part; label those as estimates and don't trust false precision. Aleatory or epistemic? Much of the uncertainty is epistemic and resolvable — customer preferences, regulatory requirements, competitive landscape can be researched — so invest in that research up to the point it stops being worth its cost, rather than gambling in curable ignorance; the residual (how the market will actually react) has aleatory and ambiguous components to be accepted, not endlessly researched. Known/unknown matrix: the known unknowns (regulations, logistics, competition) are investigable and planned for; the unknown unknowns (what could blindside us that we haven't thought of?) call for a pre-mortem and outside perspectives to surface them, and for robustness — entering in a way that is reversible and doesn't bet the company, so an unforeseen problem is survivable. Texture: if the market is fast-moving (volatility), build in agility and reserves; if it's a tangled system of local actors and rules (complexity), enter with a small probe to learn how the system responds and decentralize to local judgment; if it's genuinely novel and not understood (ambiguity), treat the entry as an experiment to build understanding rather than a bet on an understood situation. False certainty check: resist the pull (internal and organizational) to present the decision with unwarranted confidence; hold the real uncertainty honestly and decide accordingly — perhaps with a reversible pilot rather than a full committed entry. The diagnostic transforms a vague "it's uncertain whether to enter" into a structured understanding of what kinds of uncertainty are present and what to do about each — which is exactly the competence that lets the later tools be applied to the right parts in the right way.
This is the diagnostic discipline, and its value is that it comes first, before any technique. The tools of the rest of the course — expected value, decision trees, Bayesian updating, pre-mortems, decision journals — are powerful, but each applies to a specific kind of uncertainty, and applying the wrong tool to the wrong type (probabilistic precision to true uncertainty, information-gathering to irreducible randomness, confident analysis to unknown unknowns) is a characteristic and costly error. The diagnostic layer ensures the right tool meets the right uncertainty. Master it and you gain what the whole course is ultimately about: not the illusion of eliminating uncertainty, which is impossible, but the competence of handling each kind of uncertainty in the way it actually calls for — which is what good decision-making under uncertainty means.
Key Insight: The diagnostic discipline comes first, before any technique: classify the uncertainty (risk or true uncertainty; aleatory or epistemic and in what mixture; where on the known/unknown matrix; what texture — volatility, complexity, or ambiguity; and whether you're fleeing to false certainty) to determine the approach each part calls for. Its value is that each later tool applies to a specific kind of uncertainty, and applying the wrong tool to the wrong type is a characteristic costly error. Good decision-making under uncertainty is not eliminating uncertainty but handling each kind in the way it actually calls for.
This module built the diagnostic foundation of decision-making under uncertainty: the distinction between measurable risk and unmeasurable true uncertainty, between irreducible aleatory randomness and reducible epistemic ignorance, the known/unknown matrix and the danger of unknown unknowns, the textures of volatility, complexity, and ambiguity, the mind's flight to false certainty, and the discipline of matching approach to type. You can now diagnose what kind of uncertainty a decision faces, which governs every technique that follows.
The next module develops the primary toolkit for the domain of risk — uncertainty with knowable probabilities — beginning with the two concepts that underlie all quantitative decision-making under risk: probability and expected value. Expected Value & Probability builds the discipline of reasoning with the odds: how to think in probabilities rather than certainties, how to combine probabilities with outcomes to evaluate a decision, why the expected-value calculation is the foundation of rational choice under risk, and the crucial ways expected value must be adjusted for the reality of risk (variance, ruin, and the difference between money and value). Where this module taught you to recognize when you are in the domain of measurable risk, the next module gives you the tools to reason within it.