Is a coin flip really 50/50?
Everybody knows a coin toss is 50/50. It is the textbook example of a fair chance, used to start football matches and settle arguments. The interesting part is that a real coin, flipped by a real person, is very slightly not 50/50, and we know roughly by how much.
This guide covers what the research found, why long streaks are far more common than they feel, and how to get a properly fair result when it matters.
The short answer
For any practical purpose, yes. A fairly tossed coin is close enough to 50/50 that you will never notice the difference in a lifetime of deciding who goes first.
Measured over hundreds of thousands of flips, though, a small bias appears. It is not towards heads or tails. It is towards whichever side was facing up when the coin was flipped.
What the research found
In 2007 the mathematicians Persi Diaconis, Susan Holmes and Richard Montgomery analysed the physics of a tossed coin. A coin does not just spin end over end; it also wobbles, a motion called precession. Because of that wobble, the coin spends slightly more of its time in the air with its starting side facing up. Their model predicted that a coin should land the same way up as it started about 51% of the time.
That prediction was tested at scale in 2023, when a team led by František Bartoš collected 350,757 coin flips from 48 people, using coins from 46 different currencies. The coins landed on the side they started on 50.8% of the time, very close to the predicted figure. The same data showed no meaningful bias towards heads or tails in general.
The effect also varied between people. Some flippers showed almost no bias and some showed more, which fits the explanation: it depends on how much wobble a person puts on the coin.
Does a 51% bias matter?
For a single toss, hardly at all. If you do not know which side started facing up, the bias cannot help anybody, and the result is effectively 50/50. It only becomes an advantage if someone can see the starting side before they call.
There are two easy ways to remove it: have the caller choose before the coin is placed on the thumb, or use a virtual toss. The Coin Flip on this site has no physical coin and so no starting side. Each flip is an independent draw with an exact 50% chance.
Why streaks are more common than they feel
The chance of several heads in a row falls by half with each extra flip. That makes any particular streak unlikely, but streaks in general are a normal part of random sequences.
| Heads in a row | Chance | Roughly |
|---|---|---|
| 2 | 25% | 1 in 4 |
| 3 | 12.5% | 1 in 8 |
| 4 | 6.25% | 1 in 16 |
| 5 | 3.13% | 1 in 32 |
| 7 | 0.78% | 1 in 128 |
| 10 | 0.098% | 1 in 1,024 |
Those figures are for one specific run starting now. Over a long session there are many places a streak can start, so runs of five or six appear regularly. Ask people to write down a "random" sequence of heads and tails and they nearly always switch too often, because real randomness looks streakier than we expect.
The gambler's fallacy
After five heads in a row, it feels as though tails is due. It is not. The coin has no memory, and the sixth flip is 50/50 exactly like the first. The belief that past results change the odds of an independent event is called the gambler's fallacy.
What is true is the law of large numbers: over many flips, the share of heads drifts towards 50%. This happens because early streaks are diluted by the sheer number of later flips, not because the coin corrects itself.
| Flips | Typical range of heads | As a share |
|---|---|---|
| 10 | 2 to 8 | 20% to 80% |
| 100 | 40 to 60 | 40% to 60% |
| 1,000 | 469 to 531 | 46.9% to 53.1% |
| 10,000 | 4,902 to 5,098 | 49.0% to 51.0% |
Each range covers about 95% of results from a fair coin. You can watch this happen: the Coin Flip keeps a running count and percentage, and the split tightens towards 50/50 as the total grows.
How to get a fair result from an unfair coin
The mathematician John von Neumann described a neat trick for a coin you do not trust. Flip it twice. If the results differ, take the first one: heads then tails counts as heads, tails then heads counts as tails. If both flips match, ignore them and start again.
It works because, whatever the coin's bias, heads-then-tails and tails-then-heads are equally likely. The cost is that you sometimes need several rounds.
When a coin is the wrong tool
A coin only handles two options. For three or more, do not try to chain coin flips together, because it is easy to create uneven odds by accident. Use a Yes or No Wheel with a "Maybe" slice added, a Dice Roller for up to six options, or a wheel with one slice per choice.
Frequently asked questions
Is a coin flip exactly 50/50?
A physical coin is very close but not exact. Large experiments found it lands on the side that started facing up about 50.8% of the time. A virtual coin flip using a secure random generator is exactly 50/50.
Is heads more likely than tails?
No. The research found no general bias towards heads or tails. The small bias is towards whichever side was facing up before the toss.
What are the odds of 10 heads in a row?
One in 1,024, or about 0.098%. Each extra flip in the streak halves the chance.
If I get five heads in a row, is tails due next?
No. Each flip is independent, so the next one is still 50/50. Believing otherwise is known as the gambler's fallacy.