Five heads in a row does not make tails due
Heads. Heads. Heads again. A coin tool can look broken when a short history repeats. Alternating results would look tidier, but a fair independent coin does not remember what it just showed.
That matters if you use a coin to choose a first player: rerolling until the result looks balanced changes the rule you agreed on.
The next toss has no debt to repay
Under the fair independent model, each new toss has a 1/2 chance of heads and a 1/2 chance of tails. Previous heads do not increase the chance of tails. OpenStax distinguishes independence precisely by whether one event affects the probability of another.
This is an assumption about the process, not proof that every physical coin or computer implementation is fair. A suspicious tool needs its selection method examined; one dramatic-looking history is not enough.
Source: OpenStax · Independent and Mutually Exclusive Events
Be clear about which streak you mean
Exactly five specified upcoming tosses all landing heads has probability (1/2)⁵ = 1/32, or 3.125%. Five upcoming tosses all matching, whether heads or tails, has probability 2/32 = 6.25%.
Finding a run somewhere in a longer history is another question. There are several possible starting positions, and overlapping runs are related. You cannot simply treat the 1/32 calculation as the chance of seeing any run across a whole evening.
| Event | Probability |
|---|---|
| Five heads | 1/32 = 3.125% |
| Five tails | 1/32 = 3.125% |
| Either of those two | 2/32 = 6.25% |
Ten tosses do not usually produce exactly five heads
There are 1,024 equally likely ordered sequences of ten fair tosses. Only 252 contain exactly five heads, giving about 24.61%. All the other sequences have a different split.
These counts come from choosing which five of the ten positions contain heads. The calculation explains why an uneven short sample is ordinary; it does not certify any particular tool's fairness.
Keep the decision rule visible
Decide what heads and tails mean before tossing. Use one toss, or agree on best of three in advance. Avoid changing to best of five only because the first person disliked the result.
Picknado's coin history shows recent draws on the current page; it is not an independent audit record. Use the tool for everyday choices. When somebody's opportunity depends on repeated turns, a shuffled list with each person included once is a better match than repeatedly tossing for the same two names.
Sources & further reading
The links identify sources for factual statements. Schedules and worked examples are our illustrations, not user studies.
About the articles