Every strategy on this site descends from a one-page memo written in 1994 by a man who worked out of a closet.
Walter Schloss took Ben Graham's night course, worked for Graham-Newman in the early 1950s, and then ran his own partnership from 1956 to 2002 out of borrowed office space at Tweedy, Browne. No analysts. No computer. He bought hundreds of small, unloved, statistically cheap companies, priced against their assets rather than their stories, and held them until the market changed its mind.
It worked for close to half a century. In The Superinvestors of Graham-and-Doddsville (1984), Buffett published Schloss's record as exhibit one: over the 28 and a quarter years from 1956 through early 1984, the partnership compounded at 21.3% a year gross and 16.1% to limited partners, against 8.4% for the S&P over the same stretch. The longer version held up too. Per Bloomberg's obituary, from 1955 to 2002 his investments returned about 16% a year after fees against roughly 10% for the S&P 500, a figure the obituary attributes to Schloss's own estimate, which is exactly how I'll attribute it too. Buffett's summary in the 2006 Berkshire letter skips the decimals: 47 partnership years, results that "dramatically surpassed" the index, achieved with one file cabinet in 1956 and four by 2002, and a closing verdict that Walter was "one of the good guys of Wall Street."
The office situation was real, by the way. In a 1976 letter, Buffett described Walter's entire operating setup as including "a sub-lease on a portion of a closet at Tweedy, Browne," and noted that his edge showed up most in bad markets: "not only is he a man for all seasons, but... he has a special strength when facing a head wind." A Forbes profile late in his life filled in the rest of the picture: never owned a computer, got prices from the morning paper, worked 9:30 to 4:30, and summarized his whole approach in five words: "I focus on assets."
In March 1994 he compressed everything into a single typed page titled "Factors needed to make money in the stock market." Sixteen rules. They're reproduced below exactly as transcribed at rbcpa.com, original typos included, because cleaning up a primary source is how errors creep into the record. After each one, in italics: what this site actually does with it. Sometimes that's implementation. A couple of times it's honest deviation.
1. "Price is the most important factor to use in relation to value."
The entire cheapness matrix is this sentence with numbers attached. Nothing enters the portfolio on quality, story, or momentum. Price against tangible value picks the tier; everything else can only veto or nudge it.
2. "Try to establish the value of the company. Remember that a share of stock represents a part of the business and is not just a piece of paper."
Every candidate gets a full forensic write-up of the business: what it owns, what that's conservatively worth, and whether a minority shareholder will ever see it. The screen finds numbers. The write-up decides if they mean anything.
3. "Use book value as a starting point to try and establish the value of the enterprise. Be sure that debt does not equal 100% of the equity. (Capital and Surplus for the Common Stock.)"
The first axis of the matrix is price to tangible book, with adjustments disclosed line by line. And his debt warning is hard-coded twice: the leverage screen at entry, and the F-Score floor rule that rejects any company failing leverage and liquidity in the same year, no matter how well it scores elsewhere.
4. "Have patience. Stocks don't go up immediately."
Positions get roughly three years before a full re-underwrite, and the re-underwrite is a fresh look, not a forced sale. No price targets with expiration dates.
5. "Don't buy on tips for a quick move. Let the professionals do that, if they can. Don't sell on bad news."
The published rules are the only reason to buy. The thesis-break list (fraud, accounting failure, financing distress, destructive dilution, loss of the asset case) is the only news that forces a sale. An ugly headline that isn't on that list is a reason to re-check the numbers, not a reason to sell.
6. "Don't be afraid to be a loner but be sure that you are correct in your judgment. You can't be 100% certain but try to look for weaknesses in your thinking. Buy on a scale and sell on a scale up."
Looking for weaknesses is a standing institution here: the evidence page keeps a whole shelf of research against this strategy, and the rule is that if that shelf stops growing you should stop trusting this site. Buying on a scale is the add rule: additions only when the ratio is at least 15% cheaper and the book value held up. Selling on a scale up is the sell range, a band read off the stock's own history rather than one magic number.
7. "Have the courage of your convictions once you have made a decision."
Conviction here is structural, not emotional: position sizes come from the tier, and the tier comes from the grid. The rules don't get renegotiated on red days. That's what they're for.
8. "Have a philosophy of investment and try to follow it. The above is a way that I've found successful."
The Methodology page is the philosophy, in public, with the numbers. It freezes when the first live trade posts, and changes after that get dated and logged.
9. "Don't be in too much of a hurry to sell... Before selling try to reevaluate the company again and see where the stock sells in relation to its book value. Be aware of the level of the stock market. Are yield low and P/E ratios high?..."
Implemented almost literally: a position approaching the bottom of its sell range triggers a fresh review, and the range itself is recomputed against current adjusted tangible book, not the book value on purchase day. One honest deviation: I don't overlay a judgment about the overall market's level. Per-company discipline only.
10. "When buying a stock, I find it helpful to buy near the low of the past few years..."
Not a hard rule in this system, but a real bias. All else equal, I'll favor the name sitting at multi-year lows; I'm a glutton for punishment that way, so bring on the busted charts. The mechanical version: the Buy-Below price is cross-checked against the stock's own ten-year trading range, and if the grid spits out a price the stock has never actually traded at, the demonstrated range wins. Same instinct as Walter's, part temperament, part plumbing.
11. "Try to buy assets at a discount than to buy earnings. Earnings can change dramatically in a short time. Usually assets change slowly. One has to know much more about a company if one buys earnings."
The signature rule, and the reason this site is named what it's named. Assets are the first axis; earnings are the second, and only after being normalized over ten years so a single good or bad year can't drive the price. When tangible book is negative there's nothing to discount, and the write-up says so instead of switching yardsticks quietly.
12. "Listen to suggestions from people you respect. This doesn't mean you have to accept them. Remember it is your money and generally it is harder to keep money than to make it. Once you lose a lot of money it is hard to make it back."
Research tooling drafts, screens, and flags; the conclusions are owned by a human, and every material number traces to a primary filing. Nobody else's conviction, human or machine, gets to place a trade here.
13. "Try not to let your emotions affect your judgment. Fear and greed are probably the worst emotions to have in connection with the purchase and sale of stocks."
You can't delete emotions, so the design assumes them: entries, adds, and exits are all rule-driven, and the rules were published before the situations that test them. The page you're reading is part of the machinery. Public rules are harder to abandon in a drawdown than private ones.
14. "Remember the work of compounding... if you can make 12% a year and reinvest the money back, you will double your money in 6 years, taxes excluded. Remember the Rule of 72."
The quiet argument for low turnover. This portfolio's structure (hard to enter, easy to keep holding, paid while waiting via the shareholder-yield modifier) exists so compounding gets time to do the work, and so trading costs don't eat it. The transaction-cost research on the evidence page says that's where most paper strategies actually die.
15. "Prefer stocks over bonds. Bonds will limit your gains and inflation will reduce your purchasing power."
The book stays in equities. Cash and T-bill funds appear only as parking for money that hasn't found a bargain yet, and that drag gets reported as part of the record rather than edited out.
16. "Be careful of leverage. It can work against you."
He put it last so you'd remember it. It's implemented here as the one un-overridable rule in the survival screen: leverage plus deteriorating liquidity is an automatic reject, and no other strength in the score is allowed to average it away.
Where I deviate, on purpose
Honesty requires a short list of the places this is not Schloss.
He held 50 to 100 names by temperament; I cap countries, sectors, and correlated clusters explicitly and let the count fall out of the rules, which could mean anywhere from one hundred to several hundred positions. He worked almost entirely in US stocks; this portfolio screens 33 countries, which is why it carries a governance and capital-allocation gate his era mostly didn't need. He flipped through Value Line by hand; I run a screening pipeline over 120,000-plus listings. The tooling would be unrecognizable to him. The job after the screen, reading the filings one company at a time and asking what the assets are really worth, is the same job it was in the closet.
Provenance, because this site cares about that
The memo is dated March 10, 1994. No canonical scan sits behind a stable public URL, so the text above follows the rbcpa.com transcription, cross-checked against GrahamValue and MarketFolly. Transcriptions disagree in two small places: rule 14 appears as both "the work of compounding" and "the word compounding," and rule 16 as both "work against you" and "go against you." I've kept the rbcpa reading. If you have the original scan and it says otherwise, send it and I'll correct this page and say so.
Sources
Buffett, W. (1984). "The Superinvestors of Graham-and-Doddsville." Hermes, Columbia Business School. The 1956–1984 record: 21.3% gross, 16.1% net to limiteds, vs 8.4% for the S&P.
Buffett, W. (2007). Berkshire Hathaway 2006 shareholder letter, pp. on Schloss: 47 years, the file cabinets, "one of the good guys of Wall Street."
Bloomberg News (2012). "Walter Schloss, 'Superinvestor' Who Earned Buffett's Praise, Dies at 95." The ~16%/yr after fees, 1955–2002, by Schloss's estimate.
Kelly, J.R. (2021). "Walter J. Schloss: The Superinvestor of Graham-and-Doddsville." Financial History. The fullest documented account of the long record.
Forbes (2008). "Experience." The no-computer, newspaper-prices, "I focus on assets" profile.
Schloss, W. (1994). "Factors needed to make money in the stock market." The memo itself.
Schloss, W. (1998). "Sixty-Five Years on Wall Street." Talk at Grant's conference, scanned.
Nothing here is investment advice, and Schloss's results are his, not a forecast of anyone else's, least of all mine. For what the broader evidence says about strategies like his and this one, including the case against, see Does this actually work?