0.55 written beneath
Gower Street, the evening after the amendment, and the figure sits on the card in front of me: 0.88 struck through, 0.55 written beneath it in my own hand.
A number is rounded at the moment it becomes useful. That is the whole trouble. The bulletin said 0.88, and 0.88 travels well - it fits in a paragraph, it holds a headline, it lets a subscriber believe she has been handed a key. The unrounded truth was always a broader, clumsier thing, held up by two battery tiers and leaning hard on one held-out replication, and the truth did not travel because the truth took too many words.
Here is what the rounding changed for me, and I will say it plainly since no one else will: it changed my mornings. An edge at 0.88 confidence is an instruction; an edge at 0.55 is a conversation. I had been taking the first kind to the table - deciding, allocating, telling correspondents in America that the regime shift was a fact about the world. That was my benefit, and I received it in the currency of certainty, which is the pleasantest currency there is. It spends everywhere and nobody asks to see the backing.
The gap between the two figures is 0.33, and I notice that no part of that gap was produced by the market behaving strangely. The market behaved as markets do. What failed was the held-out test - the one that asks whether the pattern survives contact with days it has never seen. The L19 and L20 batteries in April did their work quietly, the way a clerk does, and the clerk’s answer was: less than you thought.
Say it plainly, then: the edge is real, era-dependent, and not reviewer-proof, and a person who needs it to be reviewer-proof should not build a household on it.
So the next action is this - I will rewrite the summary line myself tonight, before the next bulletin goes out, to read “0.55, battery tier 2 of 4,” and I will not permit the shorter version into print again, including in my own letters. The decision is made at this desk, with the lamp still low, and the old figure goes in the drawer with the other hypotheses that lived longer on paper than in the world.