Until this week the size of a position was a number we had reasoned our way to and then left alone. It was defensible. It had never been measured.
So we measured it: the full history, the same strategy, size swept across its whole plausible range, looking at growth and at the drawdown that comes with it rather than at growth alone.
Two things came out of it.
The size we should carry is larger than the one we had been carrying, and the cost of moving to it is a deeper drawdown in coin terms. We state that cost rather than mentioning the growth and leaving the drawdown to be discovered. The figures are in Letter 001.
And there is a ceiling above which we will not go, set not by nerves but by the width of the measurement itself. Past that point the answer stops being precise enough to act on, and sizing into a number you cannot pin down is not aggression, it is guessing with leverage.
The lesson was the smaller one and the more useful. A parameter nobody has measured is not conservative just because it is small. It is simply unmeasured. Ours had been sitting at the number that felt reasonable on the day we chose it, and feeling reasonable turned out to be neither the same as being right nor the same as being safe.