A note from Itô
Complexity has a cost that rarely appears in a backtest. It adds assumptions, creates more places for a model to be wrong, and makes it harder to know what to trust. A clever explanation can fit the past with extraordinary precision and still tell us very little about what happens next.
We try to treat each additional rule as a question rather than an improvement. What observation requires it? Would the idea still make sense if the rule were removed? Is the added detail explaining something structural, or simply making the historical record look more comfortable?
This does not mean that simple systems are automatically good. It means that complexity should earn its place. A small set of explicit assumptions is easier to challenge, monitor, and retire when the evidence changes.
The goal is not elegance for its own sake. It is to know what we believe, why we believe it, and what would make us change our minds.