Galton whistle rests on the laboratory bench at University College.

The Galton whistle rests on the laboratory bench at University College. This instrument, designed to test the upper limit of human hearing, produces a tone of increasing frequency as the plunger narrows the air column. We recorded the results for three hundred subjects in the summer of 1892. The distribution of audibility terminates abruptly at the threshold of the instrument’s range. This truncation is not a biological fact. It is a mechanical limit. When you see the frequency curve drop to zero at the highest graduation, you will ask if the subjects simply ceased to hear. They did not. The instrument ceased to speak.

The null hypothesis states that there is no difference in the mean auditory limit between these two groups of dockworkers. The t-test yields a p-value of 0.12. We fail to reject the null. Your immediate impulse is to claim the groups are identical. That is a logical error. A failure to reject is not a verdict of “innocent”; it is a verdict of “not proven.” The effect size is 0.15, which is small, but the sample size is insufficient to overcome the noise of the measurement.

Consider the 0.35% premium observed during equity index options expiration weeks. The null hypothesis is that this return is a product of random walk. The test rejects this null at p < 0.05. We have a directional finding. However, the instrument of our analysis - the historical window - is as truncated as the Galton whistle. We see the 0.35% because the market mechanics of gamma hedging force the signal into the audible range of our current volatility regime. Below a certain liquidity threshold, the signal vanishes. This does not mean the mechanical force is absent. It means the instrument lacks the resolution to capture the friction.

The edge of the range conceals the tail of the distribution. We are measuring the capabilities of our tools as much as the qualities of the subjects. I shall recalibrate the whistle for the next series of trials.