Coin Flip Simulator & Head/Tail Frequency Counter

100 flips, and how close the result lands to 50/50

100 flips of a fair coin produced 53 heads (53%) and 47 tails (47%) in this run -- close to, but not exactly, the theoretical 50/50 split, which is exactly what genuine randomness at this sample size should look like.

A 6-percentage-point gap (53% vs. 47%) at 100 flips is well within normal statistical variance for a fair coin -- the standard deviation for 100 fair coin flips is 5 flips, so a result within roughly 1 standard deviation of 50/50 (45-55 heads) is a completely unremarkable, expected outcome, not evidence of anything unusual.

Larger sample sizes converge more tightly toward 50/50 in percentage terms (though the absolute gap in flip count can still grow) -- this is the Law of Large Numbers in action: running 10,000 flips would very likely land noticeably closer to exactly 50.0% than 100 flips typically does.

Why this isn't a tool for proving a coin is biased

Fair coin via crypto.getRandomValues. Sample frequency is not a proof of bias. This simulates a coin with a mathematically fair 50/50 probability, using a cryptographically secure random source to guarantee that fairness in the simulation itself -- it isn't testing a real physical coin, so there's no actual bias to detect here in the first place.

If you were trying to test whether a real physical coin is biased, a single 100-flip sample (or even several) wouldn't be statistically powerful enough to detect all but fairly large biases -- rigorously testing for coin bias requires a much larger sample size and formal statistical hypothesis testing, not just eyeballing a percentage split.

The specific 53/47 result shown here is a snapshot of one run -- running the simulation again would produce a different specific split, since each run generates a fresh set of 100 independent random outcomes.

Related probability and randomness tools

For a similar convergence-toward-expectation demonstration with more possible outcomes per trial, the Dice Roller & Probability Simulator extends this same underlying randomness quality to a 6-sided (or other) die.

For generating a secure random value for an actual application need (not a probability demonstration), the Random API Key Generator uses the same cryptographic randomness source.

Frequently Asked Questions (FAQ)