Flip Coin · observed 2026-08-17
One 100-flip batch, and what it does not prove
A single run of a hundred flips will always land somewhere other than exactly fifty-fifty, so what did one dated production session actually return, and what is that count still not allowed to prove?
Pass 1
One hundred flips, read once
A bounded observation of a coin flipper starts by requesting one batch, letting it settle and recording exactly what the interface shows without adjusting anything afterward. This pass requested one hundred flips from the production tool and read the resulting tally straight from its session receipt on 2026-08-17. The bench opposite is that reading in full, with the longest streak the interface displayed during the run and the outcome of the very last flip attached alongside it.
46 heads / 54 tails
One 100-flip session
Method: production browser batch, observed 2026-08-17.
LIMIT A single batch neither proves nor disproves fairness.
Pass 2
What actually generated the flips
A tally is only as trustworthy as the randomness behind it, so this pass records what the tool itself discloses about that source rather than guessing. The production method copy states that it calls crypto.getRandomValues when the browser exposes it, and otherwise falls back to a time-mixed Math.random. Reading that disclosure is not the same as independently auditing it: this notebook observed the stated behaviour, it did not run a separate entropy test against either code path.
- Heads
- 46
- Tails
- 54
- Longest streak
- 3
- Last flip
- Heads
- Tool's disclosed source policy
- crypto.getRandomValues (fallback: time-mixed Math.random)
Pass 3
What the tool is for, and what it is not
Nothing in this notebook changes what the tool is meant to settle. It is built for casual, reversible choices and for keeping a lightweight session log of flips already made, not for situations where a certified, auditable source of randomness is required. Reading the receipt and the entropy disclosure above is enough to use it for the job it was built for; it is not enough evidence to justify using it for gambling, contractual draws or any other procedure that needs certified fairness.
Casual choices only
Recommended use boundary
Method: production decision boundary, disclosed 2026-08-17.
LIMIT Not for certified draws, gambling or official procedures.
The observed batch returned 46 heads and 54 tails. One batch cannot establish fairness or entropy quality. The tool discloses a browser-crypto path and a fallback policy, but this notebook did not verify which path produced this session or audit either generator. The receipt records this run; it is not evidence for a certified process.