"This company trades at a PER of 40—way too expensive." "That stock's PBR is 0.8, so you'd profit even at liquidation." Open any stock forum or news feed and verdicts of cheap-or-expensive, handed down from a single number, come pouring out. But a PER of 20 just means "if current earnings hold, it takes roughly 20 years to recoup your principal." Nothing more. It is not a ruling that the stock is expensive. Whether a company is cheap or dear never falls out of one number. You have to overlay six of them—market cap, EPS, PER, PBR, ROE, dividends—before even an outline appears. So let's walk through what each measures, where the optical illusions sneak in, and why, of all things, you must compare only within the same industry.

Six Rulers for Reading a Stock's Price Tag

Valuation metrics fall into two camps. Absolute ones simply gauge a company's "size" and "earning power" (market cap, EPS, ROE, dividends). Relative ones weigh those against the share price to judge "cheap or expensive" (PER, PBR). Let's lay out the whole map first.

Market Cap — A Company's True Size

A company with a 50,000-won share price versus one at 500,000 won—which is bigger? Trick question. The per-share price alone won't tell you. It swings entirely on whether the shares were split finely into many or bundled into few. Market cap is the share price multiplied by shares outstanding, showing what the market values the whole company at. And its biggest use is precisely this: stripping away the illusion of "the price of one share." To compare size, look at market cap, not the share price. Apple alone sits at a market cap of around $3.4 trillion, keeping it among the very largest single companies in the world.

One thing to nail down: market cap is not the cash sitting in the company's bank account. It is only the market's valuation; that money is not in the company's vault. In practice it's used to classify large-, mid-, and small-cap stocks, decide index inclusion, and compare company sizes.

EPS — Earnings Attributable to a Single Share

We've covered size; now for "earning power." EPS (earnings per share) divides total net income by shares outstanding, showing how much a single share earned over the year. It's the number that lets you line up the profitability of wildly different-sized companies on one scale. An EPS of $6 means one share generated $6 of net income that year.

The denominator really matters here. When shares outstanding rise (through a rights offering, say), the company can earn exactly the same and yet each share's slice shrinks. And another one—plenty of people treat EPS as cash hitting the shareholder's account. It isn't. EPS is merely accounting profit allocated per share, separate from the actual dividend. The company can split that profit out as dividends or stack it up for reinvestment. EPS serves as the denominator for PER (up next), and because share prices lurch on whether quarterly and annual EPS beats expectations, it's the most-watched number of every earnings season.

PER — How Many Years of Earnings Is the Price?

Now the main event. PER (price-to-earnings ratio) divides the share price by EPS, showing "how many multiples of the company's earnings the price I'm paying represents." Intuitively, assuming earnings hold steady, it roughly tells you how many years it takes to recoup your investment. A 20,000-won price with a 1,000-won EPS gives a PER of 20—meaning "if current earnings hold, today's price is recouped over about 20 years of earnings."

"High PER means expensive, low PER means cheap"—the most common formula there is, and it's only half right. A high PER likely means the market has already baked expectations of much higher future earnings into the price. If a company's current earnings are small but growing fast each year, investors rush to buy the larger future earnings in advance, the price forms high relative to current earnings, and the PER swells. Companies with dim prospects or heavy risk get low PERs instead. So PER is less a price tag than a thermometer of market expectations.

✍️ Editor's note — Honestly, calling a high-PER stock "expensive" was a mistake I made as a beginner too. But chewing on the formula PER = price ÷ EPS gives the answer. The numerator (price) borrows from all future expectations, while the denominator (EPS) is the last 12 months of actuals. Putting a growth stock's PER of 40 next to a value stock's PER of 12 and declaring the former "expensive" is like weighing an apple against an orange and insisting the apple is heavier. One number can't sort cheap from expensive. It only carries meaning within the same industry and against the company's own historical PER.

PER comes in flavors. Trailing PER uses the last 12 months of reported results—a confirmed figure, but backward-looking. Forward PER uses next year's expected earnings, reflecting the future but lurching entirely if that estimate misses. And one decisive limit: PER is meaningless for loss-making companies. A negative net income makes the denominator negative, and the calculation loses all meaning. That's exactly why attempts to value early-stage startups by PER so often spin uselessly.

PBR — Price Against Book Net Assets

When earnings are erratic—or outright negative—PER is useless. That's when PBR (price-to-book ratio) steps in. It divides the share price by book value per share, showing how expensively or cheaply the market values the company against the net assets on its books. Below 1 means the market price is lower than book liquidation value. A bank stock at a PBR of 0.8, for instance, could in theory return more than your investment even if the company were liquidated—though reality is never that tidy.

A common trap is the belief that a PBR under 1 is always an undervalued gem. Plenty of times the market deliberately marks it low because it thinks the assets aren't really worth their book value, or because it expects poor future profitability. PBR also discriminates by industry. It carries real weight for banks, insurers, and steelmakers heavy in tangible assets like plants and real estate, but fits poorly for IT and service firms with almost no assets. Where PER looks at earnings, PBR looks at assets—so for companies with erratic earnings, the two cover each other's blind spots.

ROE — Efficiency of Turning Capital Into Profit

ROE (return on equity) divides net income by shareholders' equity, showing how efficiently the company turned the capital it got from shareholders into profit over the year. An ROE of 20% means it made 20 won of profit from 100 won of capital in a year. Since it compresses "how well management deployed shareholders' money" into a single number, it gets cited constantly.

"Higher ROE is always better"—another trap. Heavy borrowing shrinks equity and can inflate ROE, so you have to view debt levels alongside it to see real skill. One-off gains can also make a single year spike, so reading a multi-year trend is safer. And ROE links back to the PER and PBR we just saw: companies that consistently post high ROE often earn high PER and PBR because the market expects growth. So ROE can double as the reason "why the PER is high."

Dividends — Cash That Arrives Just for Holding

Everything so far has been a tool for judging "cheap or expensive." Dividends are a different beast—cash arrives just for holding the shares. A dividend is the slice of a year's profit a company hands shareholders in cash or stock. The board sets the size, and it's paid to whoever sits on the shareholder register as of the record date. The ratio of dividend to share price is the dividend yield, a core metric for investors who prize stable cash flow. For reference, Samsung Electronics' annual common-stock dividend is reported to be around 1,444 won per share.

✍️ Editor's note — The idea that a heavy-dividend company is automatically a good company is another myth that has spread like gospel. Plenty of blue chips deliberately pay almost nothing, reinvesting profit to grow bigger. As we saw with EPS, a company can split its earnings as dividends or stack them up as reinvestment, and both can be right answers. Also, a dividend isn't a fixed promise—it shrinks or stops when results sour. Don't dive in on the dividend-yield number alone; check whether that dividend has been consistent.

Cautions and Limits — Numbers Are Questions, Not Answers

We've looked at the six separately; now for the trap underlying all of them. None alone can rule on "cheap or expensive." Here's where each one breaks down, in one table.

The most important principle is simple: compare within the same industry, and against the company's own history. Growth stocks (tech, biotech) carry high PERs, value stocks (banks, utilities, telecom) low ones, and cyclicals (steel, refining, semiconductors) throw in an extra illusion—at the cycle peak a brief earnings spike makes the PER look deceptively low. Putting a manufacturer's PER of 12 next to a software firm's PER of 40 and concluding the former is "cheap" is comparing apples and oranges. Only by overlaying the six metrics and comparing within the same industry does the illusion of a single number get filtered out.

Price Instinct Is Built by Guessing for Yourself

Understanding valuation metrics intellectually is a different thing entirely from having the instinct to gauge actual prices. Just as PER and PBR ultimately ask "where the price sits relative to the company's fundamentals," price instinct gets honed by constantly estimating and checking against the answer. In PriceGuess's daily estimation game, guess the price of a fresh target each day, and train with the higher-lower game to quickly judge which of two is pricier. The habit of overlaying multiple clues rather than ruling on a single number is the very same muscle you flex when reading stock valuation.