The story in brief
In October 1985, Evelyn Marie Adams, a convenience-store clerk from Point Pleasant Beach, New Jersey, won a $3.9 million New Jersey Lottery jackpot. Four months later, in February 1986, she won a second jackpot, this time $1.4 million.(Wikipedia contributors, 2026) News coverage at the time quoted odds of 1 in 17 trillion against any one person winning two lotteries this way — a number so large it made the story a national sensation, and later a textbook example in the statistical literature on coincidences.
Claim as commonly circulated, and what the record supports
Claim as circulated
A woman beat odds of 1 in 17 trillion — a number far larger than the world's population — making her double win one of the most statistically impossible events ever recorded, seemingly requiring an explanation beyond ordinary chance.
What the record supports
The 1-in-17-trillion figure is real, but it answers a narrower question than the one that matters: it's the probability that one specific person, buying exactly one ticket per drawing, wins both of these two particular lotteries. Statisticians Persi Diaconis and Frederick Mosteller recalculated the odds of a double winner turning up somewhere in the United States, given how many people play, how many tickets they buy, and how many lotteries run — and found the real probability of the event that actually surprised people (a double winner making news) was close to 1 in 30.(Diaconis & Mosteller, 1989)
Source trail and verification status
Adams’s win dates, prize amounts, and individual odds are well documented and uncontested.(Wikipedia contributors, 2026) The recalculated population-level probability is Diaconis and Mosteller’s own published analysis, part of their broader methodological treatment of coincidences — cited here as the primary source for the reframing.(Diaconis & Mosteller, 1989) An independent 1994 skeptics’ newsletter account reports the same two figures (1 in 17 trillion; roughly 1 in 30), corroborating that this is the standard, settled account of the case rather than a version this site constructed itself.
Worked comparison: two different questions, two very different answers
| Question asked | What it calculates | Approximate odds |
|---|---|---|
| “What are the odds that Evelyn Adams, specifically, buying one ticket per drawing, wins these two specific lotteries?” | The probability of one named individual matching two independent low-probability events | About 1 in 17 trillion |
| “What are the odds that some double lottery winner makes news in the United States within a similar few-month window, given how many people play, how many tickets they buy, and how many state lotteries run?” | The probability that at least one success occurs across a huge number of tickets, players, and drawings | About 1 in 30(Diaconis & Mosteller, 1989) |
The first number describes an event that essentially never happens. The second number describes an event that happens about once every generation or so somewhere in the country — which matches reality far better, since double lottery winners, while rare, are not unheard of.
Candidate explanations
No hidden mechanism is needed here — this is one of the clearest cases where a full accounting of “how many opportunities were there?” resolves the entire puzzle. Millions of people bought lottery tickets across dozens of states every week during this period, many buying far more than one ticket per drawing, across many drawings per year. Counting the true number of ticket-purchase “opportunities” for a double win — rather than treating Adams’s own two purchases as the only relevant chances in the universe — is exactly the law of truly large numbers at work: a per-ticket probability that looks astronomically small becomes, summed across a truly large number of tickets and players, an event you’d expect to see somewhere every so often.
What this doesn’t establish
This case does not mean lottery odds are misleading in general, or that winning any lottery once is unremarkable — a single win is exactly as rare as its stated odds say. What it shows specifically is that “the odds against this one person” and “the odds against this event happening to someone” are different questions with very different answers, and that a dramatic-sounding probability is often the answer to the narrower, less relevant question.
Related
- The law of truly large numbers covers the general principle this case is the canonical real-world illustration of.
- Multiple opportunities and selection effects covers the underlying “probability of at least one success” arithmetic in full.
- How to evaluate a coincidence claim — step 6 (choose the reference class) and step 7 (count the opportunities) are exactly what separates the two odds figures in this case.
Key takeaways
- The famous “1 in 17 trillion” figure is correct for a narrow question (this specific person, these two specific lotteries) but is not the probability of the event that actually made news.
- Recalculating for the real reference class — anyone, buying any number of tickets, across any state lottery, within a similar time window — brings the odds down to roughly 1 in 30.
- This case is the standard textbook illustration of why “count the true number of opportunities” is often the single most important step in evaluating a coincidence claim.