LotterySequential Randomness Tests
The chi-square test checks whether numbers appear the right amount of times. This checks something different — whether the order they appeared in shows any real statistical pattern, using two established methods: a runs test (do high/low sums cluster together more than chance would predict?) and lag-1 autocorrelation (does one drawing's sum predict the next one's?).
Runs Test — Clustering
Observed runs
—
—
Approximate p-value
—
—
Lag-1 Autocorrelation — Predictability
Correlation (this draw vs. the next)
—
Approximate p-value
—
—
Both tests run on the sequence of each drawing's sum (above/below the sample median, in drawn order — most recent last). The runs test asks "do high and low sums bunch together more than a coin flip would?" A p-value under 0.05 on either test would be genuine statistical evidence of a pattern in the sequence; a high p-value (the expected, and usual, result) means the order looks like what randomness actually produces. Neither test says anything about which numbers come next — see how lottery odds actually work.