For years Buffett told owners to grade him on one number: the profit earned on the owners' money. In 1982 he takes it back. Not because the number turned ugly, he insists, though it did. Because it stopped measuring what Berkshire really earns.

He knows how this looks. Managers tend to fire the scorecard at exactly the moment it starts scoring against them, and he says so himself: "Yardsticks seldom are discarded while yielding favorable readings." So he shows his work. The real problem is not the bad year. It is that an ever-larger share of Berkshire's money sits in companies it does not control, and the accounting rules ignore most of what those companies earn.

GEICO shows the absurdity. In 1982 Berkshire's books counted $3.5 million from it: the dividend. The $23 million of Berkshire's share that GEICO kept and reinvested appears nowhere. Had GEICO earned $100 million more and kept every cent, Berkshire's reported profit would not have moved a dollar. Add three more holdings like it, and the invisible earnings outgrow everything the books were allowed to show. His verdict: "accounting numbers are the beginning, not the end, of business valuation."

What should an owner watch instead? The long climb. Berkshire's net worth grew about 40% in 1982, helped along by a market suddenly fond of GEICO, and over eighteen years the owners' money behind each share has multiplied about thirty-eight times over. No victory lap follows, only a warning that the pace must slow: "Geometric progressions eventually forge their own anchors."

Whole companies were selling dear, so Berkshire bought small pieces of superb ones in the stock market instead. The sellers there, Buffett observes, are not always in their right minds. That is the whole opportunity.

This is last year's toads-and-princesses arithmetic, run from the other side of the counter. Rather than pay double for a whole company, buy tenths of wonderful ones at the going price. The market offers almost every great American business, he notes, priced daily by participants with "behavior patterns that sometimes resemble those of an army of manic-depressive lemmings." His one test for what to buy: a business whose every kept dollar becomes, in time, at least a dollar of market value.

Two cautions ride along. The approach needs a moody market: when prices race upward, as in the mania of 1972, the bargains vanish (back then Berkshire's insurers kept only about 15% of their money in stocks, against 80% now). And the reward keeps no schedule. Some years the market will mark good businesses down, and Buffett's answer is to buy more at the better price.

The proof is personal. His two largest paper gains, The Washington Post and GEICO, are businesses he first brushed against at ages 13 and 20 and came back to as an investor in the 1970s. His deadpan conclusion: "nostalgia should be weighted heavily in stock selection."

Here is the actual list. The whole basket cost $424 million and was worth $945.6 million by year-end, more than double what was paid, with GEICO alone nearly a third of the worth.

Two newcomers join: Time, and Crum & Forster (a parking place for cash, Buffett notes, not a keeper). And unlike a year ago, when two holdings sat in red below their cost, not one on the 1982 list was worth less than Berkshire paid.

The proudest achievement Buffett reports for 1982 is a purchase that did not happen. A large deal he had firmly committed to collapsed for reasons outside his control, and he counts the escape among the year's real accomplishments.

Buffett confesses that he "left the room once too often last year and almost starred in the Acquisition Follies of 1982." The big deal collapsed on its own, sparing him time, energy, and "a most uncertain payoff," and he suggests the fitting illustration himself: "two blank pages depicting this blown deal would be the appropriate centerfold." For the second year running, Berkshire's biggest acquisition news was a deal it did not do.

Watching the deals others did close, he feels "not envy, but relief that we were non-participants." In deal fever, "managerial intellect wilted in competition with managerial adrenaline." Pascal supplied the cure centuries ago: the ability to stay quietly in one room. And price is unforgiving. Paid too high, it "can undo the effects of a subsequent decade of favorable business developments."

The letter's most famous stretch is about what a company pays with. Cash is simple. But a company that pays with its own undervalued shares is quietly handing over two dollars of value to get one.

His rule takes one sentence: "we will not issue shares unless we receive as much intrinsic business value as we give." Why would anyone "issue dollar bills in exchange for fifty-cent pieces?" Yet managers do it constantly, because a chief's appetite for deals outruns his cash, a hunger Buffett admits to sharing, and often at just the moment his own stock is cheap. The bankers cheering him on are paid by the deal ("Don't ask the barber whether you need a haircut").

A favorite excuse is that the company must grow. Grow whom, Buffett asks. Fold your family's 120-acre farm into an "equal partnership" with a neighbor's 60, with you as managing partner, and you now run 180 acres, but your family owns a quarter less of land and crops than before. "Managers who want to expand their domain at the expense of owners might better consider a career in government."

Only three honest exits exist, and Exhibit 10 holds them: trade true value for true value, pay with stock only when the market prices it at full worth, or buy back afterward every share you issued. Beneath all three sits a plea for honest language. "Company A to Acquire Company B," Buffett says, is better read as "Part of A sold to acquire B," because what you give away matters exactly as much as what you get.

Berkshire's biggest business is insurance, and Buffett uses 1982 to explain, at last, why the industry keeps sliding into losses. The reason is structural: insurance is a commodity, and the old rules that once protected its prices are gone for good.

The forecast is blunt: "underwriting results in 1983 will not be a sight for the squeamish." By 1982 the industry was paying out nearly $1.10 for every premium dollar it took in, and even that, he warns, is a "best case," padded by soft accounting ("It's difficult for an empty sack to stand upright").

The disease has two parts. The product is identical from every seller (customers buy candy bars by brand, but nobody orders coffee with "C & H sugar, please"). And the supply is bottomless, since new capacity takes nothing but "capital plus an underwriter's willingness to sign his name." Identical product, endless supply: prices can only grind down.

Then why was underwriting profitable for decades? Regulation. Giant A legally charged what Giant B charged, rate-cutting was forbidden, and insurers could "price their way to profitability" no matter the surplus. "That day is gone." New sellers wield price as a weapon, customers have learned that insurance "is no longer a one-price business," and no tidy "cycle" is coming to the rescue: supply here is "mental rather than physical," and only a true shock (a "megadisaster") will make underwriters put their pens down.

Berkshire had no immunity: its own underwriting slipped from well above the industry average to modestly below it. The exceptions teach the lesson. Milt Thornton and Floyd Taylor kept underwriting at a profit, as they have every year since joining, and GEICO wins even in a price war because of "a wide and sustainable cost advantage," the one moat a commodity business can have.

Near its end the letter carries a want-ad: exactly what Berkshire wants to buy, printed for anyone to read, with an answer promised in about five minutes. The lesson is the clarity itself.

The six gates are in Exhibit 14, and the flavor is pure Buffett: no projections ("future projections are of little interest to us"), no turnarounds, nothing complicated ("if there's lots of technology, we won't understand it"), no hostile deals, and no haggling in the dark. In return he promises confidentiality and "a very fast answer as to possible interest (customarily within five minutes)." Cash preferred. Stock only on the honest terms of the last lesson.

The same spirit runs the head office. "In a characteristically rash move," Buffett reports, Berkshire expanded it by 252 square feet. And the letter's warmest passage is a goodbye. Phil Liesche and Ben Rosner retired after running their businesses "with every bit of the care and drive" of full owners, no rules required. "Their good character became our good fortune."