Eleven Intervals by the Lamp

Sit down by the lamp and look at this column with me before you trust it. The ledger from the Harley Street inquiries, spring of 1860 - I keep it here for teaching - holds a fortnight of figures where the deaths in one ward fell on nearly every second day. Nearly. Eleven of fourteen intervals, regular as a pulse, and for an hour I believed I had found the poisoned drain. The four gaps were nothing. I called them noise. That is the seduction: a pattern that is almost there invites you to forgive the parts that break it, and the forgiveness feels like judgement when it is wishfulness. Your question, the one every student asks at this point, is how much regularity is enough. The honest answer is that regularity is the wrong test entirely. Ask instead what would have produced the pattern if the cause you suspect did not exist. Eleven of fourteen is also what a fair coin gives you eleven times in fourteen, more often than your instincts guess. I did not find a drain. I found a coin.

Now to the bulletin before us. The claim is a return premium clustered around the settling of options contracts - 0.35 per cent, they say, persisting across the calendar. The pattern is almost there. It repeats often enough to comfort, and the mechanism proposed - hedging flows, mechanical, structural - sounds the way a bad drain sounds: plausible, sanitary, actionable. But audit the collection instrument first, as we stated before we analysed. Did the measurement change when the contracts changed? Does the count of these settling periods survive when the calendar is shuffled, the way I shuffled my ward days and watched my neat alternation collapse into scatter? If the premium vanishes under rearrangement, the finding was the coin again, wearing finer clothes.

The evidence level, if it holds, justifies a small trade written in pencil. It does not justify a doctrine, and it does not justify the chart on the committee wall.

Here is my correction, plainly. When I began this note I believed the fault lay in the gaps - that the four irregular intervals were the problem to be explained. I now believe the opposite. The gaps are the honest part of the data. The seduction lives in the eleven that fit. Trust the pattern-breakers first; they are the ones telling you the truth about the rest.