A real coin is not quite fair
Persi Diaconis and colleagues showed that a physical flip caught in the hand lands on the face it started on about 51% of the time — the coin precesses rather than rotating cleanly. Spinning a coin on a table is considerably more biased, because the heavier face tends to end downwards.
A digital flip has no physics to be biased by. Each result here comes from the system’s cryptographic random source, which is closer to a fair coin than a coin is.
Streaks are normal
In a hundred flips, a run of six or seven of the same result is likely rather than remarkable. People asked to fake a random sequence almost never write long enough runs, which is exactly how fabricated data is spotted.
Flipping many at once here shows the tally, which makes the point better than any explanation: the proportion settles towards half while the runs keep happening.
Frequently asked questions
Is a digital flip fairer than a real coin?
Yes, slightly. A caught physical flip favours the starting face about 51% of the time, and a spun coin is worse. Software has no physics to bias it.
I got seven heads in a row. Is it broken?
No — that is expected in a run of a hundred flips. Genuinely random sequences contain longer streaks than people expect.
Can I flip many at once?
Yes, with the running tally shown — which is the quickest way to see the proportion settle while streaks keep occurring.
Something wrong with this tool, or an idea for it? Tell us