Probability-Weighted Value: Bull, Base and Bear in One Number

A single fair value hides the thing that matters most: how wrong you could be, and what being wrong costs.

Ask a DCF what a company is worth and it answers with a point — precise, confident and silent about its own fragility. Risk-adjusted valuation refuses the point. It prices the same company under three explicit futures — bear, base and bull — then blends the answers by how likely each future is. The question changes from "what is this worth?" to the sharper one professionals actually trade on: what is this worth if I am wrong, and how much does being wrong cost me?

The method's signature discipline is that risk enters twice. In a genuine downturn, cash flow slows and the return investors demand rises at the same time — earnings fall while risk is repriced. A bear case that only trims growth understates the damage by half. So the scenarios move both dials together: bear takes growth down five points and the required return up one and a half; bull adds four points of growth and concedes one point of return. Each case then runs the identical machinery — a ten-year projection with growth fading to a 2.5% terminal rate — so the only differences in the outputs are the judgements, never the arithmetic.

Worked example: Example Packaging Machinery Co.

A hypothetical industrial-machinery maker (every company in this series is invented for teaching). The reported facts never change between scenarios: last year's free cash flow of $600M, net debt of $900M, 200M shares. Only the futures differ:

ScenarioGrowthRequired returnValue per share
Bear2%11.0%$30.96
Base7%9.5%$48.33
Bull11%8.5%$68.44

Weight the futures — bear 30%, and the remainder split 60/40 between base and bull (42% and 28%) — and the blend is:

(0.30 × $30.96) + (0.42 × $48.33) + (0.28 × $68.44) = $48.75

The market price is $44.00: roughly 10% below the weighted value, and — the reading this model exists to give — 42% above the bear value. The downside case is not priced in; if the bear future arrives, the holder loses roughly 30% from here. The weighted midpoint says "modestly cheap"; the bear anchor says "cheap, with a real bill attached if you're wrong." Both statements are the analysis.

Read the spread before the midpoint

The bull case is 2.21 times the bear case. That ratio is a diagnostic in its own right:

Bull ÷ bearReadingPractical response
Under 1.5xPredictable businessThe midpoint is usable
1.5x to 3xNormal uncertaintyAnchor on the bear case
Over 3xOutcome is a coin flipTreat the midpoint as noise

A wide spread means the answer depends almost entirely on assumptions — so position size should shrink as the spread widens, not grow with the midpoint's optimism. Above 3x, the weighted average is an average of two different companies and deserves no confidence at all.

Stress the probabilities, not only the inputs

Analysts love stressing growth by a point. The probabilities move the answer more, and they are where motivated reasoning hides — so set them before you see the weighted answer, and set them from the business's cyclicality and its history of missing, not from the conclusion you want. The same three cases, re-weighted:

Bear probabilityWeighted valueChange
15%$52.56+7.8%
30%$48.75base
45%$44.94−7.8%
60%$41.12−15.6%

Pessimism alone, with every cash-flow assumption untouched, moves fair value by double digits. A stable utility might honestly carry a 10–15% bear weight; a cyclical industrial or a recent IPO may deserve 35% or more. Using the same odds for every company is the quietest way to lie with this model.

When scenario weighting breaks down — and the neighbouring tools

The machinery needs positive cash flow to discount. A company burning cash has no base to project; its real question is funding risk, which belongs to the dilution radar, not a scenario engine. Binary outcomes break it differently: a failed drug trial or a lost licence is not a slower version of the base case but a different company entirely, and smoothly reducing growth does not describe it. And the macro regime is its own scenario dimension — the macro-adjusted DCF rebuilds the discount rate for the rate cycle, while rate sensitivity measures how much of a fair value is really a duration bet.

Place the blend among its siblings before acting. The reverse DCF inverts the exercise — what future does the price already assume, and what probability is the market silently assigning it? The single-case DCF is the base case with the other two futures amputated. The multiples frame the same uncertainty cruder but faster: P/E, P/S and EV/EBITDA for the market's current odds, relative value for the blend across methods, EV/FCF for the cash the scenarios project. Downside-first thinking has its own instruments: the Graham number for a margin-of-safety floor, liquidation value for the terminal downside, NAV per share for the asset anchor, and the financial metrics for the leverage that decides whether a bear case is survivable. Every input traces to the filings in the equity research terminal.

Try it yourself

QuantaVal is a quantitative valuation engine — 88 models in the Quantitative Valuation Engine (free calculator: 62) — and its Quantamental Investing AI explains the numbers in plain language. The free risk-adjusted calculator is one of the 62 free calculators: free to use immediately, no credit card — visitors get 3 shared calculations a day, resetting at midnight UTC, and the full methodology is free to read. Screen for candidates in the free screener (2,000+ stocks), price their three futures — and set your probabilities before the answer can talk you out of them.

Example Packaging Machinery Co. is hypothetical and used for teaching only. Figures are illustrative, not a forecast for any real company, and nothing here is investment advice.