In this article, we examine how a competition to find the world’s most accurate clock might be run. How could the winning clock be identified if it outperforms every existing standard for timing? The intuitive view on time-keeping is that a good clock is one that keeps time consistently and hence agrees with other clocks. This view, we argue, is mistaken. Measurement is fundamentally about making high-quality predictions. Accordingly, the goal is not consistency, but independence between the clock and its environment. We propose that, counter-intuitively, the best clock is the one that ticks most unpredictably, making its predictions the most difficult to beat. The organizers of the clock competition should award the prize to the clock that ticks most randomly.