Named by Albert Goldman (Lindy's Law, 1964), restated by Benoit Mandelbrot (The Fractal Geometry of Nature, 1982), and extended by Nassim Nicholas Taleb (Antifragile, 2012).
Estimates the remaining life of a non-perishable thing, such as an idea, method, or technology, from how long it has already lasted.
The longer a method has lasted, the longer it is likely to last
Each line is the chance that a method is still in wide use after a number of years, read from its age alone: something that has lasted A years has even odds of lasting A more, and a chance of A over A plus T of lasting T more. Set the age of the method you are judging and the years you need it for. The rule fits ideas, formats, and practices with no built-in end; it does not fit people, machines, or anything that wears out. Illustrative.
Yours: chance still in use, percentFive-year-old framework500-year-old practice
An example
A manufacturer's controller had to choose how ten years of reconciliation records would be kept, since the auditors and the tax rules required them to stay readable for a decade. A software vendor was pitching its analytics workbook format, five years old, with good search and charts built in. The alternative was a plain spreadsheet file format that had been in wide use for about forty years.
The vendor's demo was better, and the team leaned toward it. The controller asked a narrower question: which of the two was likely still to be readable when an auditor asked in year ten? Judged by age alone, the five-year-old format had about a one-in-three chance of lasting another ten years; the forty-year-old format about four in five. Nothing else on the table was known as well as their ages.
She split the decision. The archive of record went into the old format, and the vendor's tool was licensed to read those files for analysis. If the vendor folded or changed its format, the records would still open.
Why it works
For things that do not wear out, such as ideas, formats, methods, and practices, survival is itself evidence. Each year something stays in use, it has passed another round of competition, changing needs, and new alternatives, and the ones that were going to fail early have mostly failed already. So among things of this kind, the ones that have lasted longest tend to have the most life left, and a thing's age becomes a usable forecast of its future when little else is known.
The rule is useful because it needs no theory of why something lasts. It favors the old over the new in exactly the cases where the new looks better on features and nothing can yet show it will last. It does not apply to anything with a natural lifespan, which ages toward its end.
The thinking behind it
Albert Goldman Albert Goldman, an American academic and writer on popular culture, named the idea in Lindy's Law, an essay in The New Republic of June 13, 1964. He reported the folklore of the comedians who gathered at Lindy's delicatessen in New York, who held that a comedian's future run could be predicted from his exposure so far. The rule began as talk about show business careers; its later authors turned it into a general claim about how long things last.
Benoit Mandelbrot Benoit Mandelbrot, the founder of fractal geometry, reworked the folklore in The Fractal Geometry of Nature (1982), giving the Lindy name to the rule that some things' expected future life grows with their past life. His version applied to things bounded by a producer's life. His earlier study of cotton prices, The Variation of Certain Speculative Prices (1963), had shown that price changes follow heavy-tailed distributions, with far more extreme moves than the normal curve allows.
Nassim Nicholas Taleb Nassim Nicholas Taleb, a former options trader who writes on risk, extended the rule in Antifragile: Things That Gain from Disorder (2012) to anything nonperishable without a natural limit, such as ideas, books, technologies, and practices: the life expectancy of such things grows with their age. This broader form is the one now called the Lindy effect. His books are essays for a general reader, with technical appendices, and some statisticians have faulted his treatment of statistics.
How to run it
- Check that it is nonperishable. Apply the rule only to ideas, methods, formats, and practices, never to people, machines, or companies that age.
- Date each option. Find how long each has been in wide use, counting from when it spread, not when it was invented.
- Set the horizon. Write down how many years the choice has to hold: a retention rule, a contract term, a planning cycle.
- Compare the odds. Use age over age plus horizon as each option's rough chance of lasting the horizon.
- Weigh it against the features. Decide whether the newer option's advantages are worth its lower odds of being around.
- Hedge the young choice. Where the new option wins, keep an exit: an export path, an old-format copy, a second supplier.
Where it goes wrong
It gets applied to things that wear out. A forty-year-old machine is not likely to run forty more; spot the error when the subject has parts, a body, or a balance sheet. Use failure rates and condition data for those.
Age is measured from the wrong date. A method invented long ago but used widely only for three years is young. Count from when it spread.
The rule blocks every new idea. Taken as a ban, it would have kept everyone on the old format forever. Use it to price the risk of the new and to demand an exit, and choose the new where its advantages outweigh that risk.
When to use something else
- Reference Class Forecasting. When there is a record of similar cases to forecast from, beyond a single thing's age.
- Barbell Strategy. When the choice is how much to stake on new, uncertain options against safe, proven ones.
- Technology S-Curve. When the question is whether a technology is running out of improvement and a successor is rising.
Where it sits in SOLVE eX
Go to the source
- Lindy's Law, The New Republic, 1964
- The Fractal Geometry of Nature, 1982
- Antifragile, 2012