You can run every scorecard, pass every check, and still lose money. The numbers tell you what a business has done. They cannot tell you what you have misunderstood, or where they quietly stop applying. This chapter is the judgement that sits on top of the numbers, and the quiet mistakes that cost careful people the most.
Buy what you know
Source: Lynch, chs. 1-2
Peter Lynch's most-quoted line is also his most misunderstood. "Buy what you know" does not mean "buy a company because you use its products." It means start with industries and companies you actually understand, and then do the financial work.
Buy what you know does not mean buy what you use. It means stay where you can actually judge.
The edge in starting inside what Lynch called your circle of competence is that you notice things early. You see that a store is busier this season, that a product is catching on, or that a rival is stumbling, often well before it shows up in the official filings. From there the job is the same as for any other stock. Read the statements, judge the edge, estimate the value, and insist on a margin of safety. This is the same edge over Wall Street from Chapter 1, turned from an advantage into a discipline.
Six kinds of stock
Source: Lynch, chs. 6-10
Lynch sorts stocks into six types. You buy each type for a different reason and judge it by different numbers. Mixing them up is one of the most common mistakes in investing.
- Slow growers. Mature companies growing 1% to 3% a year, like utilities and old telecoms. You buy them for their dividends and steadiness, not for a rising price. Sell if the dividend looks at risk.
- Stalwarts. Large, established companies growing 10% to 12% a year, like Coca-Cola and Procter & Gamble in their prime. You buy them for steady compounding. Sell when the price gets silly, and otherwise hold for years.
- Fast growers. Smaller companies growing more than 20% a year. The highest possible return and the highest risk of overpaying. Watch for the growth rate to slow, which is the most dangerous moment in their life.
- Cyclicals. Companies whose profits rise and fall with the economy, like carmakers, airlines, and steel. Oddly, you buy them when their P/E looks high, because profits are at a low, and sell when their P/E looks low, because profits are at a peak.
- Turnarounds. Troubled companies that may or may not recover. Most fail, but the ones that make it can return many times your money. Only try one with real evidence that a recovery is already underway, not on hope.
- Asset plays. Companies sitting on assets, like real estate, cash, or patents, worth more than the whole company's market price. Often dull on the surface, which is the point. The market has not noticed what is underneath.
What to avoid
Source: Graham, Buffett, Lynch
- Trying to predict where the price goes next. Nobody does it reliably, and the effort steals time from work that actually pays off.
- Trading on chart patterns. Moving averages and the like tell you nothing about what a company is worth.
- Catching a falling knife. A stock down 50% is not cheap by default. It may be down because the business is broken.
- Selling great companies just to spread your money around. If you find ten companies you understand and trust, owning all ten is fine. Scattering across forty random names you barely know is not diversifying, it is drifting.
- Watching the portfolio every day. At that timescale the noise drowns out the signal. Read the annual report when it lands, and ignore the ticker in between.
- Confusing a bumpy ride with real risk. A price that bounces around is the cost of higher returns over time. The real risk is losing your money for good, which is a different thing.
Limits of the tools
A few things the scorecards and tools on this site are not, on purpose:
- Not a buy-or-sell machine. There are no buy or sell calls anywhere. The scorecard lays out the numbers. You do the judging.
- Not one-size-fits-all. The cutoffs were set by Buffett and Clark mostly on consumer brands and industrial companies. Banks, insurers, and property companies keep their books differently and do not fit cleanly. Tech companies often fail the research-spending mark without that being bad. Read the colors as a starting point, not a verdict.
- Not focused on a single year. Wherever we can, we use five-year averages and trends. A single great year can lie. A company that hit a 30% return on equity last year but averaged 12% over a decade is not a 30% company.
- Not a replacement for the story. The numbers cannot tell you whether the edge is widening or fading, whether management can handle a big change, or whether a new technology is about to make the product cheap and common. As Damodaran puts it, value is a story plus numbers, never numbers alone.
- Built and tested on limited samples. A return backtest like the one behind Piotroski's F-Score leans on databases that survivors dominate, so a real downturn can be worse than the test suggests. The Altman Z-Score has a different limit: it was fitted in 1968 on a small set of manufacturers, the bankrupt ones included (that was the whole point, to tell them apart), so it travels poorly to other industries and eras rather than to survivorship. Treat both as rough flags, not laws.
Use the tools to organize your thinking and to flag obvious patterns of quality, price, and risk. Use the statement tabs on a ticker page to check the numbers and trace the story. And see the sources for the original books, which nothing here can replace.