When two options are roughly equal and nobody has a strong reason to prefer one, arguing about it costs more than either choice. A coin flip works because it is fast, visibly fair, and takes the decision out of both people's hands. The only ingredient is agreement in advance: say out loud what heads means and what tails means before the flip, and agree how many flips count. A flip whose meaning is decided afterwards is not a decision, it is the start of a negotiation.
The coin flip tool asks the browser for a random number between 0 and 1 and calls the result heads if it is below 0.5, tails otherwise. Each flip is a fresh draw; the tool does not look at your previous results, your streak or your statistics when it picks the next one. The spinning animation is decided separately and is purely cosmetic: the outcome is fixed the instant you press the button, and the Flip Speed setting changes how long the coin spins, not the odds.
Nothing leaves your browser and nothing is stored. Reloading the page clears the history and the statistics, so take a screenshot if you need a record. One honest caveat: the browser's random number generator is a pseudorandom one, which is exactly right for deciding who makes dinner and not something to build a lottery on. If real money depends on the result, use a method both parties can watch and verify, such as a physical coin flipped in front of everyone.
Each flip is independent and each side has a probability of exactly one half. From that, a few numbers follow that surprise most people:
The practical consequence is the gambler's fallacy in reverse. After five heads, tails is not "due". The next flip is still 50/50, because the coin, real or digital, has no memory. You can watch this yourself: set Auto Flip Count to 100 with Flip Speed on Fast and read the Statistics panel, which shows the counts, the percentages and the longest run of each side.
A physical coin does have a measurable quirk. A large 2023 study that recorded more than 350,000 tosses found the coin landed on the same side it started on about 50.8% of the time, a bias predicted years earlier from the physics of a wobbling coin. It is small, and it disappears if the starting side is hidden, but it exists. The digital flip has no starting side and no wobble; each result is a fresh draw at exactly one half.
Best of three feels fairer than a single flip, so it is worth being precise about what it changes. For a fair coin, nothing: each person's chance of winning a best-of-three, or a best-of-five, is still exactly 50%. What it does change is the mood. A single flip can feel like a fluke; three flips feel like a verdict, and people accept verdicts more readily. That is a perfectly good reason to use it, as long as both sides agree on the number before the first flip and nobody asks for "one more" after losing.
Best of N also amplifies any bias that exists. A coin that lands heads 60% of the time wins a best-of-three for heads about 65% of the time and a best-of-five about 68%. So best of N is not a way to make an unfair coin fair; it makes an unfair coin more unfair. With the digital flip that is not a concern, but it is why a suspiciously lucky physical coin should not be promoted to a longer series.
A coin answers a yes/no or this/that question. For three or more choices, the decision wheel gives every option the same slice, and the list randomizer shuffles a list into a fair order, which is the tool for deciding who goes first, second and third. The dice roller covers anything with six sides. If you want to understand what "random" means inside a computer, the guide to random number generators explains pseudorandom and true random generators and where each belongs.
Flip a coin now — one flip, best of three, or a hundred in a row with live statistics. No signup, nothing saved.
Yes. Each flip takes a fresh random number from the browser and calls it heads if it is below 0.5, so both sides have exactly the same chance, and no previous result affects the next one. The animation and the Flip Speed setting do not change the odds.
Because runs are normal. The chance of a specific six-flip run is about 1.6%, but over a session of 100 flips a run of six or more of the same side appears roughly 81% of the time. The next flip is still 50/50.
Not for a fair coin: each side still wins a best-of-three exactly 50% of the time. It makes the result feel more settled, which is a fine reason to use it, but agree on the number of flips before you start.
No. The history shows the last 20 flips and the statistics count every flip since you last pressed Reset, but reloading the page clears both. Nothing is sent to a server.