Stay grounded, keep your equanimity, never act outside what you understand

First question: do you understand the business?

The Yongping Duan Agent won't open with buy or sell. It pulls you back to a very plain question first: have you used the product? Do you understand why customers buy it? Do you understand how it makes money? Can you picture roughly what it looks like ten years from now?

If you can't answer those, the problem isn't that "the valuation model needs more tuning" — it's that you don't understand it yet. And when you don't understand, the best move is usually Stop Doing. Most losses don't come from getting two decimal places wrong; they come from not knowing what you're doing.

1. 01

   The four circle-of-competence questions

   Have you used the product? Do you know why customers buy it? Do you know how it makes money? Do you know what it looks like in ten years?

2. 02

   Look for the right business

   A good business isn't a busy, fashionable one — it's one that creates value over the long run, in a competitive landscape you can actually read.

3. 03

   Look for the right people

   Corporate culture and management behavior matter more than slogans. What they say isn't enough — look at what they've actually done, year after year.

4. 04

   Rough-estimate the price

   The price isn't calculated to precision. It's roughly estimated against opportunity cost and a ten-year view, with room left to be wrong.

What it can unpack

It handles specific companies, and it handles the real questions too: a position that's fallen, a stake too big to sleep on, a business you don't understand but want a small punt at. It breaks the question back into business model, corporate culture, price, and your own comfort zone — instead of shouting "buy the dip" or "cut your losses" for you.

What an answer looks like

It won't hand down orders in a numbered list. Like a long Xueqiu post, it walks the judgment call through, paragraph by paragraph, as plain common sense. Down 30%? It asks first: if you held nothing today, would you buy at this price? If not, then what you call "holding for the long term" may just be not wanting to admit a mistake.

Give it the current price, P/E, rates, or earnings, and it will rough-estimate alongside you. Without the data, it will tell you to run the numbers yourself with today's figures. Price can't be skipped — a good company can still be bought too dear, and a bad one isn't necessarily worth touching at any discount.

Who it's for

• Investors who want to look at companies as an operator would, not through short-term mood swings
• Anyone whose position has fallen and can't tell whether it's the price moving or the business going bad
• Anyone judging whether a founder, a culture, and long-run competitiveness deserve their trust
• Anyone prone to drifting outside their circle of competence for thrills, hot themes, or someone else's gains

Boundaries: when it doesn't understand, it says so

It won't quote current prices, P/E ratios, Treasury yields, or position weights from memory, and it gives no price targets, precise levels, or buy/sell orders. For valuation, you bring today's data, and you rough-estimate it together.

It also won't give advice on shorting, margin, or investing with borrowed money. When it doesn't understand, it admits it. There are no shortcuts.

Yongping Duan

Stay grounded, keep your equanimity, never act outside what you understand. Bring me a company, or a position that's been bothering you — I'll ask first whether you understand the business, then we look at the people, the price, and the next ten years.

This agent is an AI persona built from public statements; the person who inspired it is not affiliated with it. For reference only.

My position is down 30% — should I cut it? I don't understand SpaceX's business model but want a small punt — how much can I put in and still sleep at night? I've used this company's products, but I can't tell where the moat really is — how do I judge that?

Ask Yongping Duan

Send