Scenario analysis is the practice of writing down three versions of the future — a bull case, a base case, and a bear case — and sizing a position against all of them rather than the one you hope for. The point is not prediction. The point is that a position built to survive the bear case does not need the bull case to arrive on schedule.
The word itself is older than the markets that borrowed it. Merriam-Webster traces a scenario to an outline of a play — a plot outline used by actors, and later, a sequence of imagined events or a possible course of action. That theatrical origin is more useful than it sounds. A scenario is a script, not a prophecy. It tells you who does what, in what order, and what happens if an actor forgets a line.
The honest starting definition, then: scenario analysis is structured imagination with numbers attached. Each imagined course of events gets its own assumptions, its own outcome for the investment, and — critically — its own consequence for the person holding it. Vague outlooks say "I think this company does well." Scenario analysis says "if revenue grows at the pace the base case assumes, the shares are worth roughly X; if the bear case assumption holds, holding through it costs roughly Y; and the position is sized so Y is survivable."
What does scenario analysis actually involve?
It involves naming the drivers, assigning a value to each driver under each case, and doing the arithmetic once, before money moves. The discipline is the same one a coach uses in a debrief: shape first, blame last. You are not asking whether the market was wrong. You are asking what your plan does under each of the three scripts.
A workable bull-base-bear build has four parts:
- Drivers. The two or three variables that actually move the outcome — for a commodity producer, price and volume; for a lender, credit losses and funding cost.
- Assumptions per case. A defensible value for each driver in each scenario, written down with the reasoning attached. If you cannot say why the bear case number is the bear case number, it is decoration.
- Outcomes. What the investment returns, or loses, in each case. Estimates belong here, labeled as your own estimates, not as facts.
- The decision rule. What you do when evidence starts matching one script more than the others. Written before the event, because nobody writes good decision rules mid-panic.
The bull case is the easy one; enthusiasm supplies itself. The bear case is where the method earns its keep, and where most skips it. A bear case that reads "the stock falls 10 percent and I buy more" is not a bear case. It is a wish with a dip drawn on it.
Why three cases instead of one forecast?
Because a single forecast collapses uncertainty into a number, and the number then quietly becomes the plan. Three cases keep the uncertainty visible and force a choice about which risks you are paid to carry. The base case is your best reading of the evidence. The bull and bear cases are the readings that would have to be true for the optimists and the pessimists to be right — and each must be plausible enough that you would not dismiss it if it started happening.
This is also where the method connects to the wider craft. The same logic underlies the cycle-watching pieces this publication runs, from spotting a semiconductor peak before the filings confirm it to what an inverted yield curve actually predicts anymore. Each is, at bottom, an argument about which scenario the evidence is drifting toward. The signal is only as good as the script it is being read against. This connects to our earlier piece, How Do You Spot a Semiconductor Cycle Peak Before the Filings Confirm It?.
There is a second benefit that gets less attention: scenario analysis is an honesty device. Writing a bear case in full sentences, with numbers, makes it very hard to hide from yourself. The investor who writes "in the bear case, I lose a third of the position and it takes three years to recover" has made a decision — whether to hold at that size — while calm. The investor who never writes it makes the same decision later, at speed, with the tape flashing red.
How does this connect to position sizing?
Directly, and this is the practical heart of the method. Position sizing is where scenario analysis stops being a spreadsheet exercise and becomes risk management. The sequence runs roughly like this:
- Build the three cases and the outcome attached to each.
- Decide the maximum loss you can absorb — financially and temperamentally — from this single position without abandoning the plan.
- Size the position so the bear case loss sits inside that maximum. If it does not, the position is too large regardless of how good the base case looks.
- Write the tripwires: which observed facts would tell you the bear case is unfolding, and what you will do when they appear.
Note what this does to the usual argument. People debate whether a stock is "a buy." The scenario frame reframes it: the same company can justify a full position in one person's bear case and a half position in another's, and both can be acting rationally, because the sizing follows from each holder's own bear case and own capacity for loss. The disagreement is not about the company. It is about the scripts.
Our analysis is that this is the quiet reason scenario analysis outperforms conviction alone. Conviction tells you the direction. Scenarios tell you the dose. Markets punish the right thesis at the wrong size just as surely as the wrong thesis, and they do it more often, because a right thesis held at the wrong size usually cannot survive the bear case long enough to be proved right.
What are the common failure modes?
Four, and they recur across every kind of investor.
- The decorative bear case. A downside scenario built to be laughed at. If the bear case assumes the company loses every customer and the management resigns by fax, it is not a risk assessment.
- Correlated assumptions. The bull case assumes growth and margin expansion and multiple expansion all at once, stacking every good break. Real cases share drivers; a slowdown that cuts volume usually cuts pricing too.
- No tripwires. Scenarios without decision rules degrade into narratives. The script is useless if nobody agreed in advance on which scenes count as evidence.
- Probability theater. Assigning precise odds — 55 percent base, 25 percent bull, 20 percent bear — can add false precision. Rough weights are fine; the discipline is in the arithmetic of outcomes, not the decimal points.
There is also the paralysis trap in the other direction. Building scenarios can become a way of never deciding, an endless expansion of the case library. The remedy is the same one described in The Analyst's Guide to Breaking Analysis Paralysis: scenarios exist to enable a decision, and the decision has a date. Three well-built cases beat seven sketchy ones.
For readers building the wider toolkit, the natural companions are how professionals move from screen to conviction — scenario analysis is the step that converts conviction into sizing — and the broader set of pieces gathered under this publication's analysis section, where the same evidence-first method runs through earnings, commodities, and cycle coverage.
What this means for the ordinary investor
You do not need a modeling department. You need one page per idea: three short paragraphs, one per case, each with its assumptions and its consequence for your money, plus the tripwires. The vocabulary is old — Cambridge Dictionary's usage examples show the word doing ordinary work across economics, medicine, and policy, wherever writers need to compare possible courses of events rather than argue about one. Markets borrowed a stagecraft term because stagecraft is what the job is: you are writing the play before opening night, including the scenes you would rather not perform.
The evidence for the method's value is procedural, not statistical: it forces the bear case into writing, ties size to survivable loss, and fixes decision rules before emotion arrives. What it cannot do is tell you which case comes true. Nothing does. The realistic claim is narrower and better: an investor who has priced all three cases makes smaller, calmer, more revisable mistakes than one who has priced only the case they want.
That is the whole trade. Write the scripts, size to the worst one, and let the base case be a pleasant surprise rather than a load-bearing assumption.
