The Exponential Decay Function: How Ebbinghaus Quantified the Rate at Which Unrehearsed Information Is Lost

In 1879, inside a quiet room in Berlin, a 29-year-old German psychologist named Hermann Ebbinghaus sat to the rhythm of a ticking metronome, reading aloud from a list of 2,300 artificially constructed three-letter words. He was attempting to measure something invisible: the exact speed at which the human brain discards new information.[2]

The actionable takeaway from his years of self-experimentation is absolute: without active rehearsal, you will lose exactly half of what you just learned within the first hour. To retain information permanently—whether a new language, a medical protocol, or a software framework—you must intercept the memory just as it begins to degrade, a technique now coded into every major spaced-repetition learning platform.[3][6]

Before Ebbinghaus, memory was a philosophical concept. He turned it into a mathematical function. To isolate pure retention, he had to eliminate the advantage of prior knowledge. He created "nonsense syllables"—consonant-vowel-consonant combinations like "WID" or "ZOF" that carried no meaning and triggered no prior associations in the brain.[2][3]

Over the course of a year, Ebbinghaus memorised lists of these syllables, waited for a specific interval, and then measured how many repetitions it took to relearn the list perfectly. He tested intervals of 20 minutes, one hour, nine hours, one day, two days, six days, and 31 days.[2]

The results, published in his 1885 monograph Über das Gedächtnis (On Memory), revealed a brutal biological default. The forgetting curve is not linear; it is an exponential decay function. The steepest drop occurs almost immediately after learning ceases.[2][3]

Within just 20 minutes of perfect memorisation, Ebbinghaus found his retention had plummeted to 58.2 percent. After one hour, it sat at 44.2 percent. By the 24-hour mark, two-thirds of the unrehearsed information was gone, leaving a retention rate of just 33.7 percent.[2]

He quantified this loss with a formula: R = e^(-t/S), where R is memory retention, t is time, and S is the relative strength of memory. The equation dictates that as time increases, retention drops exponentially, but the curve eventually flattens out. Whatever survives past the first few days tends to remain accessible for a month.[3]

For 130 years, the Ebbinghaus curve was a foundational textbook concept, yet it had rarely been subjected to a rigorous, exact replication. That changed in 2015, when researchers Jaap Murre and Joeri Dros at the University of Amsterdam recreated the original protocol using modern experimental controls.[1]

The 2015 replication yielded a decay curve nearly identical to the 1885 original. "We successfully replicated Ebbinghaus’ forgetting curve," Murre and Dros wrote, noting that their subject retained 58 percent of the syllables at the 20-minute mark, perfectly mirroring the 19th-century baseline.[1][6]

But the decay rate is not fixed for all types of data. The variable S in the Ebbinghaus equation—memory strength—is heavily influenced by how meaningful the information is to the learner. Rote memorisation of isolated facts decays much faster than interconnected concepts.[3][4]

A study published through the University of Memphis Digital Commons explored this by comparing the forgetting rates of paired associates. Researchers found that when learners attach semantic meaning to a pairing, the initial decay velocity slows significantly compared to the pure rote memorisation Ebbinghaus tested.[4]

What actually happens to the "forgotten" information? According to an analysis in ResearchGate on the form of the forgetting curve, the data is not necessarily erased from the physical brain structure. Instead, the retrieval pathways degrade, making the information inaccessible to conscious thought.[5]

Ebbinghaus proved this through his concept of "savings." Even when he could not consciously recall a single syllable from a list after 31 days, it took him significantly fewer repetitions to relearn the list than it did to learn it the first time. The brain retains a latent trace of the exposure.[2][5]

The antidote to the exponential decay function is spaced repetition. By reviewing the material at precisely the moment the curve dips—say, at 24 hours, then three days, then a week—the learner forces the mathematical curve to reset to 100 percent.[3][6]

With each subsequent review, the decay rate slows. The curve flattens. After four or five spaced rehearsals, the retention line becomes nearly horizontal, locking the information into long-term declarative memory with minimal ongoing effort.[3]

Today, this 1885 mathematical function powers a multibillion-dollar educational technology industry. Algorithms in platforms like Anki, Duolingo, and SuperMemo calculate the exact time (t) at which a user is statistically likely to forget a specific flashcard, prompting a review just before the memory trace vanishes.[6]

The human brain is an aggressive editor, optimized to discard the irrelevant to make room for the necessary. Ebbinghaus did not just discover that we forget; he proved that forgetting is a predictable, quantifiable mechanism—one that can be systematically bypassed by anyone willing to do the math.[2][6]