A fair coin has an exact 50% probability of landing heads and 50% landing tails on any single flip — but flip it 10 times and you'll rarely see exactly 5 heads and 5 tails. That's not a flaw in the coin; it's a completely normal, expected feature of how randomness actually behaves in the short term.
The 50/50 probability describes what happens on average over a very large number of flips, not what any specific small sample will look like. This is the law of large numbers in action: the actual ratio of heads to tails only converges close to 50/50 as the number of flips grows very large — a handful of flips will naturally show real variation, sometimes quite a bit of it.
How Coin Flip Probability Is Calculated
Probability of heads (or tails) on any single flip = 0.5 (50%)
Probability of a specific sequence of n flips (like heads, heads, heads) = 0.5^n
Expected number of heads over n flips = n × 0.5, though the actual observed number will typically vary somewhat from this expected value, especially for small n.
A Worked Example
Flipping a coin 10 times, the "expected" result is 5 heads and 5 tails — but getting exactly that split has only about a 24.6% probability. Results like 6 heads/4 tails or 4 heads/6 tails are each individually about as likely, and even 7/3 or 3/7 splits happen a meaningful percentage of the time purely by chance. Flip 1,000 times instead, and the ratio reliably lands much closer to 50/50 — the law of large numbers smoothing out the short-run variation that's completely normal at 10 flips.
Common Mistakes to Avoid
- Believing a coin is "due" for tails after several heads in a row: each flip is independent — previous results have zero influence on the next flip's probability (this misconception is called the gambler's fallacy).
- Expecting a small sample to closely match the theoretical probability: variance is completely normal and expected in small samples — it's not evidence the coin is unfair.
- Confusing "50/50 chance" with "will happen exactly half the time": probability describes long-run tendency, not a guarantee for any specific number of trials.
- Assuming a real coin is perfectly fair: physical coins can have very slight biases from asymmetric wear or minting, though the effect is generally too small to matter for everyday use.
Bottom Line
50/50 is a long-run average, not a short-term guarantee. Use a Coin Flip Simulator to flip any number of times and see how the actual heads/tails ratio behaves — and how it gets closer to 50/50 as the flip count grows.