Ask anyone to fake a coin sequence and they betray themselves the same way: they alternate too often. A real coin repeats itself in long streaks, but a streak feels non-random, so people switch sides too soon. The runs test counts the blocks of repeats and compares them to chance — too many blocks and the meter catches the forgery.
Chance is streakier than intuition. In sixty fair flips, the longest run you should expect is around six in a row — a length most people would never dare to write down.
The same blind spot lives in space. A truly random scatter of stars is clumpy: it has knots and empty voids, because each point ignores the others. A grid that has been gently nudged looks random but is too even — and it is the fake. We call clumps "patterns," but clusters are exactly what randomness looks like.
That reflex — reading meaning into noise — is apophenia, and it is the engine behind lucky streaks, hot hands, and constellations. The bench doesn't lecture; it measures.
The bench stalled mid-run. The lesson continues without it — everything you have already done is safe.