Are Lottery Draws Actually Random? A χ² Reality Check
We run a uniformity test on every game we cover. Here is how the chi-squared method works, and what it says about whether draws are fair.
It is one thing to assert a lottery is fair; it is another to measure it. For every game in the warehouse, PatternSight runs a chi-squared uniformity test on the observed number frequencies and stores the result. The method is simple enough to explain in a paragraph.
The test in plain English
If a game is fair, every number should appear about equally often over the long run. Chi-squared compares what actually happened to that even expectation, squares the differences so overs and unders do not cancel, and sums them. A large total means the observed pattern is hard to explain by chance; a small total means the bumps are ordinary variation. We convert that total into a p-value — the probability of seeing deviations this large in a genuinely fair game.
Observed counts over the analysed window. A uniform game trends flat; bars are descriptive, not predictive.
What we find
For well-run national games, the p-values are consistent with fair draws — the deviations are within what randomness produces. That is the boring, correct result. Where a game shows a surprising value, it is far more often a data-quality artifact (a matrix change, a feed glitch) than a real bias, and the windowing PatternSight applies is designed to keep the test honest across a game's history.
The point is not to "beat" the test — it is to report a game's behaviour truthfully. A fair game tested fairly should look fair.