Bear case

Now the most important hour of the entire deep dive, and the one amateurs skip. After two days building the case for a business you chose to study because something attracted you, your model is a monument to your own thesis. Confirmation bias (M6.04) guarantees it: you have been unconsciously weighting confirming evidence and explaining away the rest for 48 hours. The single most powerful corrective ever devised is to argue the other side better than the people who believe it. Not to skim the bear case and rebut it, but to rebuild it stronger than its own authors made it, and then see whether your thesis survives.

Why is steelmanning the bear case worth an hour?

Why it is the highest-value hour. Three reasons. First, it is the cheapest possible way to find the flaw in your thesis before the market charges you tuition for it, a pre-mortem on the analysis itself. Second, if your thesis survives a genuinely strong bear case, your conviction is now earned rather than merely felt, which is exactly what lets you hold through the drawdown that tests every thesis (the behavioral edge, see Day 3 in context). Third, and this is the subtle one, the exercise converts a vague cloud of "risks" into a short list of specific, falsifiable variables, which is what the pre-mortem and the sell criteria are built from. You cannot monitor "the moat might erode"; you can monitor "studded ratio falls below X for two straight years." Steelmanning is the machine that turns worries into triggers.

Where can you find the bear case?

Where to find the bear case (free sources). The bear is usually already written, often better than you would write it:

• Short-seller reports: Hindenburg, Muddy Waters, Kerrisdale, Grizzly, Viceroy, Bear Cave, Culper (US); for India, Hindenburg's Adani work and the standing archive of SEBI orders and forensic newsletters. Read them for method, not verdict: even a wrong or self-interested short report often names the real fragilities.
• The 10-K/AR's own Risk Factors: read the specific ones, not the boilerplate (M8.01); the company's lawyers have listed the bear case for you.
• Value Investors Club (free with a delay), SumZero, Seeking Alpha bear articles, FinTwit/X, and Reddit threads. Mine for the strongest single argument, ignore the price-target noise.
• Competitors' calls and decks: a rival explaining why it is winning is your target's bear case in disguise.

How do you build a bear case?

The method (four steps).

1. Collect every bear point: write them all down, however weak, from all sources.
2. Rebuild the strongest version. Strip the ad-hominem and the obviously wrong; add the data the bears missed; make each surviving point quantitative. Your job is to make the bear case something you would be embarrassed to have missed: the version a smart, hostile portfolio manager would put to you.
3. Convert to break-variables. For each strong bear point, write the falsifiable form: "My thesis is wrong if ______," with a monitoring metric and a threshold. This is the deliverable.
4. Classify the bear. Is it a valuation bear (the business is fine, the price is too high, so you can wait for a better entry, the thesis is not broken) or a thesis bear (the business itself is impaired: fatal, the moat or the runway is gone)? The distinction governs the action: a valuation bear says "watch and wait"; a thesis bear says "there is no price at which the story is what you thought."

Worked example: a Titan bear case

Worked steelman (Titan). The strongest bear case, rebuilt to be genuinely uncomfortable:

• Lab-grown diamonds (LGD): a thesis-level threat to the profit engine. The steelman: natural-diamond prices have fallen hard as LGDs became indistinguishable to consumers and collapsed in price; the studded (diamond-jewellery) segment is Titan's margin engine, richer than plain gold. If LGDs commoditize diamonds, the studded margin premium compresses structurally — not cyclically. Weak bear version: "LGDs will kill Titan." Strong bear version, quantified: if studded gross margin compresses by N points and studded mix stalls, group EBIT margin falls from ~12–13% toward ~10–11%, and since the reverse DCF requires margins to hold, a chunk of the premium valuation is unwarranted. Break-variable: studded ratio and studded segment margin, tracked over 2+ years; the natural-vs-LGD price spread.
• Valuation leaves no margin of safety: a valuation bear. The steelman: even granting a great business, the price implies mid-teens growth and >25% ROCE for a decade-plus; any stumble de-rates a high multiple violently. This is a valuation bear, not a thesis bear: it argues about price and entry, not about whether Titan is a good business. Break-variable: the reverse-DCF implied expectations vs delivery; the price you'd need for a margin of safety.
• Competition and formalization cutting both ways. The steelman: Reliance Jewels, an aggressively expanding Kalyan, Malabar, and regional players are also chasing the organized-share shift; the thesis assumes Titan keeps its share of the formalizing market, but formalization could accrue partly to well-capitalized rivals, forcing margin-dilutive competition. Break-variable: Titan's EBIT margin and its share of organized jewellery (not total) over time; competitor store-add pace.
• Gold-price and macro sensitivity. The steelman: big-ticket discretionary jewellery demand is sensitive to gold-price shocks and slowdowns; a sharp move can create both a demand air-pocket and (absent perfect hedging) an inventory/margin hit. Break-variable: same-store growth and buyer-count growth through a gold spike; hedging-disclosure quality.

What does the steelman conclude?

Now the honest verdict, which is what steelmanning is for: the LGD point is the one genuine thesis-level risk (it attacks the margin engine, not just the price), and it is the variable your pre-mortem and sell criteria must center on; the valuation point is real but is a wait signal, not a broken signal; competition and gold are manageable, monitorable risks. A thesis that has survived this steelman is one you can hold with earned conviction — and one you now know exactly how to falsify.

Compare this to the amateur who read one bull deck and bought: you have done the opposite of fall in love; you have interrogated.