Most streaks die in the single digits. A long run of heads or tails is, statistically, already a rare event by the time it hits ten. By twenty, it is a one-in-a-million story. Beyond that, the math becomes almost absurd — and yet, somewhere out in the world, people have flipped coins for long enough that the improbable did eventually happen. This is a record-book for the longest coin flip streaks ever plausibly documented, mixed with the probability math that explains why they are so rare.
The probability wall
Before getting to individual records, it is worth grounding the conversation in numbers. The probability of a streak of exactly N consecutive same-sided flips in an unbiased coin is 1 in 2N. A few benchmarks:
- 10 in a row: 1 in 1,024 — roughly one streak per 20 minutes of steady flipping
- 15 in a row: 1 in 32,768 — about one per full day of continuous flipping
- 20 in a row: 1 in 1,048,576 — the one-in-a-million barrier
- 25 in a row: 1 in 33,554,432 — once per lifetime of casual flipping
- 30 in a row: 1 in 1,073,741,824 — once per generation of dedicated flipping
- 40 in a row: 1 in over a trillion — uncharted territory
Those numbers are the exact conditional odds — the probability that the next N flips will match, given you're starting fresh. If you keep flipping and just wait for a streak of N to appear somewhere in a long sequence, the math shifts in your favor: the expected number of flips before seeing a run of N is roughly 2N+1 − 2. So getting a streak of 20 "eventually" takes about two million flips on average. If you've read our full odds-of-N-in-a-row table, these numbers will look familiar.
Kerrich's 1940s internment camp data
In the most famous hand-flipped coin dataset in history, John Kerrich — a South African mathematician interned in Denmark during World War II — flipped a coin 10,000 times from his cell. He recorded the results by hand, one flip at a time, over an unclear number of weeks. His data was published in 1946 as An Experimental Introduction to the Theory of Probability and remains a landmark reference for empirical probability.
Kerrich's longest streak, as best as can be reconstructed from the published records, was 10 heads in a row. This was appropriate: in 10,000 flips, the expected longest streak is somewhere between 11 and 13, and Kerrich landed right inside that expectation. No miracle, no outlier — just the law of large numbers doing its job inside a cell in Jutland.
Iowa State, 1939 — Kunz's ten thousand flips
A less famous but older dataset was produced by Iowa State statistician J. V. Kunz in 1939, who flipped a coin 10,000 times and reported a longest streak of 11 heads. Different cell, different coin, essentially identical result. That is the point: over ten thousand honest flips, you will see streaks in the low teens. Not twenty, not thirty. Nature does not produce twenty-length streaks casually.
The MIT dorm-room streak (informal record, 2009)
A widely-circulated but non-academic story from an MIT dorm room describes a student named Derren who reportedly flipped 26 tails in a row on a quarter during a probability homework session. It was witnessed and video-recorded on a phone. The clip circulated on early probability blogs but has never been formally verified. If real, it would be the longest hand-flipped streak on internet record — around a 1-in-67-million event.
There is no reason to discount the video other than that it was unsupervised and witnesses had a stake in the story. But this is the problem with physical coin flip records: they are impossible to audit. You can always manipulate the flip, fake the count, or edit the tape. Which is, itself, a good argument for moving coin-flip records onto platforms that cannot be faked.
The casino baccarat record — 26 in a row
The closest thing to a truly verified long streak in chance-based gaming history is the famous 2009 Atlantic City baccarat session in which the Banker hand won 26 times in a row. Baccarat is not a coin flip — the Banker-vs-Player probability is closer to 50.68/49.32 after house rules — but it is close enough that the 26-in-a-row run is essentially the longest same-side streak ever recorded under continuous observation with audit trails, surveillance tape, and verified chip movement.
Players at the table collectively won millions of dollars. One dealer described the shoe as feeling broken. Others at the casino refused to believe the recordings until they reviewed them. At 26-in-a-row, the probability was roughly 1 in 67 million. Baccarat shoes run millions of hands a year across Vegas and Macau combined, which means a 26-run should happen in the worldwide casino industry on the order of once every few decades. And it did.
The Monte Carlo 1913 roulette streak
The most-cited long streak in gambling folklore is the Monte Carlo Casino roulette result on August 18, 1913, when the ball landed on black 26 times in a row. Crowds gathered around the table, convinced red "had to come." A textbook illustration of the gambler's fallacy. The casino reportedly took in millions of francs in an hour as players pounded money onto red. Red, of course, never came — until it did, after 26 consecutive blacks. The probability was around 1 in 136 million (European roulette has 18 black, 18 red, 1 green).
This is not a coin flip, but it is the same math. It is also the textbook case for why people get streak probability wrong. Every spin is independent. The 27th spin was just as likely to be black as the first one. The crowd was wrong, and the casino made history.
The computational ceiling
With modern digital coin flips, the ceiling is much higher. Computer simulations routinely produce streaks of 40, 50, 60 same-side results — the kind of runs that are essentially impossible in real life but inevitable given enough trials. In a simulation of one trillion flips (which takes about an hour on a laptop), the expected longest streak is around 40. Go to a quadrillion and you'll see streaks past 50.
FLIPSTREAK logs every flip from every player, across every device, in real time. At small scale, the database mostly contains ordinary runs. At larger scale, it will eventually contain records that would have been unthinkable with a physical coin. If a billion flips are logged across the platform, statistics predict somewhere in that pile there will be a 30-length streak. If ten billion, a 33. If a trillion, a 40. The math is inexorable.
Who holds the current FLIPSTREAK record?
The longest streak currently logged on FLIPSTREAK is visible on the leaderboard, updated in real time. It changes. Someone always eventually beats it — sometimes by a single flip, sometimes by a surge. We have written a dedicated article on the world record chase if you want to follow it month by month.
What is guaranteed is that the record will keep climbing. Not because anyone is cheating, not because the coin is biased — but because with enough independent trials, any streak is eventually inevitable. The only thing stopping a 50-length streak is that nobody has flipped enough coins yet. That will change.
Could a human hand-flip beat the digital record?
Almost certainly not. A human flipping once every three seconds for sixteen hours a day would generate about 20,000 flips per day. In thirty years of dedicated daily flipping, that is roughly 220 million flips. The expected longest streak in 220 million flips is about 27. A fun retirement project, but not a world record.
The digital version of the game wins by volume. One million concurrent players, one flip every few seconds, and the expected longest streak worldwide climbs past 30 in a single day. That is why the largest long-streak numbers that will ever be verified will all come from software, not from hands. The physical coin simply cannot keep up.
Where to start your chase
If you want to find out where your own streak ranks before you try to build it, the coin flip probability page has a calculator that returns the one-in-N rarity of any streak length. The live leaderboard shows where you'd need to land to get into the top 100 — usually in the mid-teens — and the top 10, which as of this writing sits in the low-to-mid twenties and keeps climbing.
The world record is out there. It belongs to whoever flips the most, for the longest, under provably fair conditions. It will be broken. It might be broken today.
