How to build a bull case and a bear case for any stock

Professional analysts rarely write "buy" without also writing down what would make them wrong. The bull case and the bear case are two arguments built from the same facts. The quality of your decision depends on how honestly you build the one you do not like.

Why both sides matter

When we like a stock, we notice good news and explain away bad news. This is confirmation bias, and it affects experienced investors as much as beginners. Writing the opposing case forces you to look at the evidence you were skipping.

It also gives you a plan. The bear case tells you which facts to watch. If they start to appear, you already know what they mean.

The template

For each side, answer the same five questions:

Fill both columns with the same effort. If one column takes five minutes and the other thirty seconds, the short one is the one you need to work on.

Building the bull case

Look for evidence, not adjectives:

• Growth drivers: new products, markets, pricing power, backlog.
• Margin path: operating leverage, cost programmes, mix shift.
• Catalysts: earnings, launches, index inclusion, buybacks.
• Trend: price above rising long-term averages, higher lows, volume on up days.

Building the bear case

The strongest bear cases usually come from four places:

• Valuation: what growth does the price assume, and how often has the company delivered that?
• Concentration: dependence on one customer, product, supplier or regulator.
• Quality of earnings: profits growing faster than free cash flow, rising receivables, heavy stock-based pay.
• Positioning: crowded trades, extreme sentiment, a price far above its moving averages.

A worked example (illustrative)

Imagine a hardware company that has doubled in a year on AI demand.

Bull: a large backlog of AI server orders, improving gross margin, strong free cash flow, clear uptrend.

Bear: the stock trades far above its 200-day average after a vertical run, the backlog is concentrated in a few hyperscaler customers, and margins in server hardware have historically been thin and competitive.

The honest verdict might be that both are right: the business is improving and the price already reflects much of it. That leads to a useful conclusion that is neither buy nor sell: good company, poor entry, wait for a better risk/reward or a confirmed earnings date.

That is exactly the kind of Neutral call a structured debate produces when the evidence is mixed. It looks less exciting than a bold call. It is usually more useful.

Questions that expose a weak thesis

• Could someone who disagrees with me use the same three facts?
• Am I arguing against the strongest version of the other side, or a straw man?
• If the stock fell 20 % tomorrow with no news, would I buy more or panic?
• Which single data point would make me exit?

Let a machine argue with you

The bear case is hard to write for a stock you love, because you are not neutral. That is one of the best uses of AI in investing: it has no position, so it can build the other side without flinching.

Kairon does this with two separate agents, a bull researcher and a bear researcher, who rebut each other before a third agent decides. You can read the whole exchange. Try it on the stock you are most confident about.

This guide is educational and not investment advice.