Methodology — Base-Rate Expectations (Mauboussin)

The feature a black-box DCF can't give you: not "−68%, expensive," but "the price requires ~35%/yr; only 1 in N companies that size ever did that." Display-only — it never feeds a fair value, so it cannot inflate one (the cardinal rule holds by construction).

This is the public methodology. The published source and every step of the read are named; the exact interpolation scheme and the frequency cutoffs behind each verdict are proprietary calibration and deliberately not enumerated.

The idea

Reverse the DCF to recover the growth the current price embeds, then look that growth up in an empirical base-rate table — the historical frequency with which companies of that size actually sustained it.

Source. Michael J. Mauboussin, Dan Callard & Darius Majd, The Base Rate Book, Credit Suisse Global Financial Strategies (Sep 2016), Sales-Growth chapter ("Base Rates by Decile, 1950–2015"). The universe is the top ~1,000 global companies by market cap since 1950 (all sectors, includes dead/merged firms). CAGRs are real (inflation-adjusted to 2015 USD).

The computation

Implied growth. Take the reverse-DCF stage-1 growth the price requires (see Fair Value).
Real terms. Deflate today's revenue to real-2015 USD (to pick the size bucket) and the required growth to real terms (to look up the table), via the standard Fisher relation.
Size bucket. Place the firm in the published revenue-size buckets.
Survival lookup. Convert the required growth into an exceedance frequency — the historical share of same-size companies whose long-run growth met or beat the requirement — interpolated from the published exhibit's frequencies.
Verdict. The frequency maps onto a five-notch scale, from most to least commonly achieved: UNDEMANDING → PLAUSIBLE → AGGRESSIVE → HEROIC → NEVER-OBSERVED (no company of that size in ~65 years of data ever delivered it). The cutoffs between notches are fixed calibration parameters, disclosed directionally here and stable across the universe so ranks are comparable.

Credibility — name-relative, not an absolute cliff

The reverse-DCF implied growth is a revenue-growth requirement only under the engine's held-margin assumptions. For a genuine grower it's a credible top-line expectation; for a stable compounder (e.g. a name growing 2% whose price "requires" 32%) it is largely a held-margin / buyback / multiple artifact. We therefore compute a continuous coverage read — how much of the required growth the firm's own trailing growth already covers — not a fixed threshold — and phrase the read accordingly: a high-coverage name reads "already covering ~N% of the bar," a low-coverage name reads "on fundamentals alone the price implies X%; the balance is the multiple/buyback/quality premium, not growth."

Honest limits

Descriptive, not a forecast. Growth barely autocorrelates; the base rate says how rare the priced-in outcome has been, not what the firm will do. A demanding name can keep compounding on margin, buybacks, or a quality multiple the DCF doesn't credit.
Survivorship biases the frequencies up (each N-year cell only includes firms that survived N years) — the safe direction for a HEROIC read (if anything conservative).
Published-bin granularity means an arbitrary required growth is interpolated; a 0-cell means "none observed in ~65 years," not "impossible."
The Expectations page ranks the whole universe by this verdict.