Yhprum’s law: why some systems work Yhprums Lawwhen they shouldn’t
Some things work for reasons nobody designed into them.
Along the highway to Seling, about 65 kilometres from Aizawl in Mizoram, farmers set up small bamboo stalls stocked with vegetables, fruit and sometimes dried fish. Nobody sits at them. A price is chalked on a board, and a box sits next to the produce. You take what you want, drop the money in the box and pick out your own change if you need it.
Mizos call these nghah lou dawr, shops without shopkeepers.
On paper, the plan should fail. Anyone could walk off with the vegetables, or the box. The farmers are away in their fields. Yet the stalls have run for years, and the owners told The Better India in 2016 that their customers “have never failed us”.
That gap between what should happen and what does has a name: Yhprum’s law.
Yhprum’s law is the idea that systems which should not work sometimes do. “Yhprum” is “Murphy” spelled backwards, and the law is the optimistic mirror of Murphy’s law. Its short popular form is “everything that can work, will work”. Two economists who studied eBay, Paul Resnick and Richard Zeckhauser, put it more carefully: systems that shouldn’t work sometimes do.
Where did Yhprum’s law come from
The name is older than its best-known use. The earliest known use points to Alan Abelson, a columnist at the American financial weekly Barron’s, in his column of 9 December 1974. A chemical engineering newsletter quoted his version in April 1975: “Anything that should go wrong won’t.” Abelson was poking fun at gloomy economic forecasts he thought were overblown. The column itself is hard to trace, so the attribution rests on that 1975 quotation.
Abelson’s line is a different idea from the modern one. It is a joke about prophets of doom. The version most people meet today is the opposite of Murphy’s law, and it is often credited to Zeckhauser, a Harvard economist. He did not coin the name. He and Resnick gave it the form worth remembering.
The modern form came out of research on eBay. From its early years, eBay let buyers and sellers rate each other after every sale. Resnick and Zeckhauser studied its 1999 records and found something textbook economics said should not happen.
Rating a seller takes time, earns you nothing and mostly helps strangers. A purely self-interested buyer would skip it. Yet buyers left ratings more than half the time, and almost all of them were positive. Sellers received a negative rating only about 1% of the time.
A system built on unpaid, anonymous and nearly always polite reviews looked useless. Yet every week hundreds of thousands of second-hand items changed hands between people who would never meet. The researchers’ explanation was courtesy. People leave a good rating after a smooth sale the way they say thank you.
In 2002, Resnick and Zeckhauser worked with a postcard dealer, John Swanson, and a researcher, Kate Lockwood, to test what a good reputation was worth. Swanson sold matched batches of old postcards under his well-rated account and under brand-new ones. Their 2002 paper closed by saying that the eBay Feedback Forum illustrates Yhprum’s law.
Yhprum’s law examples, in India and beyond
In December 2005, the journal Nature sent science entries from Wikipedia and Encyclopaedia Britannica to experts without telling them which was which. Across 42 usable reviews, the experts found 162 errors, omissions or misleading statements in Wikipedia and 123 in Britannica. That works out to about four per article against three. Serious errors came to four on each side.
Britannica rejected the findings in March 2006 in a rebuttal titled “Fatally Flawed”, and Nature stood by its study. Take Britannica’s side if you like. An encyclopaedia anyone could edit, written by unpaid volunteers, still came far closer to the professionals than its design suggested it should.
Mumbai offers a better example. The city’s dabbawalas collect home-cooked lunches from homes and deliver them to offices, then bring the empty tins back the same day. Stefan Thomke of Harvard Business School, who wrote a case study on them, described about 5,000 workers with little formal schooling delivering upwards of 130,000 lunches, six days a week, by train, bicycle and handcart, without mobile phones. He found mistakes extremely rare.
The error rate is where the story gets exaggerated. The figure of one mistake in 16 million deliveries, often labelled “Six Sigma”, traces back to Forbes magazine around 1998 and was never a formal certification. Thomke’s case describes the performance as Six Sigma equivalent or better. Treat the exact number with caution. The reliability that impressed Thomke is not in dispute.
Each of these systems failed a test someone had set in advance. The eBay prediction assumed buyers who never do anything unpaid. Wikipedia lets anyone, vandals included, change almost any page. The dabbawalas fail almost every item on a modern logistics checklist, from software to staff qualifications.
What the tests left out was how people behave inside a system they belong to. Courtesy on eBay. Volunteers watching the pages they care about on Wikipedia. On Mumbai’s trains, a coded marking on each tin and a team whose members know their routes and one another. When a system “shouldn’t work” and does, the prediction usually carried a narrow picture of people.
You meet this closer to home. The office process nobody ever wrote down, run for years by one careful clerk. The housing society where maintenance gets paid on time with no penalty clause. If you have seen one of these, you have seen Yhprum’s law at small scale.
Is Yhprum’s law true, or just wishful thinking
It is a proverb, as Murphy’s law is. Neither predicts anything. The useful question is when Yhprum’s law helps and when it misleads, and history has a hard answer to the second.
Before the space shuttle Challenger broke apart on 28 January 1986, the rubber O-rings sealing its booster joints had shown erosion on earlier flights. Those flights landed safely, and managers read each safe landing as evidence the seals were good enough. The physicist Richard Feynman, a member of the presidential commission that investigated the disaster, attacked that reasoning in his appendix to its June 1986 report. Erosion, he wrote, was a warning that the equipment was not behaving as designed. He compared it to Russian roulette, where a safe first shot “is little comfort for the next”.
Challenger is Yhprum’s law turned into a trap. Something worked when the design said it shouldn’t, and people counted that as proof instead of treating it as a question.
The eBay study carries its own warning. In the postcard test, new accounts given one or two deliberate bad ratings still sold at normal prices. Resnick and colleagues suspected that most buyers never clicked through to read negative ratings and relied on the single overall score. A rating system that works partly because nobody reads it closely will keep working only until a dishonest seller notices.
There is also a counting problem. People remember eBay and Wikipedia. The rating schemes and volunteer projects that collapsed as predicted stopped being talked about. Yhprum’s law describes the survivors.
How to use Yhprum’s law at work and at home
When something works that you expected to fail, find out why before you change it. A messy arrangement often runs on a habit or a person that no redesign will notice until they are gone.
Write down the reason it works in one sentence. If you cannot, treat the success as luck that has not yet run out, and keep checking.
Watch the small deviations: the late delivery, the error caught by chance. Feynman’s point was that these are the early reports from a system telling you it is not behaving as designed.
And give people a little more credit than your model does. Resnick and Zeckhauser’s buyers left feedback nobody paid them to leave. Design with that in mind, then build checks for the few who will not.
A system that works for a reason you understand is an asset. One that works for a reason you do not understand is on loan.
Daily Reflection
Name one thing at work or at home that runs better than it should. Ask what keeps it running, and what would happen if that person or habit disappeared.
