The separation between data and strategy is especially important because precision can create false confidence. A database can tell you exactly what happened, while the investment process still makes a poor decision by weighting the wrong factors, ignoring differences between business models or market environments, or sizing the position badly. Being able to distinguish a data failure from a thesis, portfolio-construction, or ordinary uncertainty failure is what makes the process genuinely improvable.
Exactly. Good data improves the visibility of the system, but it does not automatically improve the decisions made from it.
A useful investment process should be able to trace each bad outcome back to the correct layer: inaccurate data, inappropriate classification, weak factor weights, poor portfolio construction, or simply unavoidable uncertainty.
Without that separation, every failure gets blamed on “the model,” and the process never learns what actually needs fixing.
The separation between data and strategy is especially important because precision can create false confidence. A database can tell you exactly what happened, while the investment process still makes a poor decision by weighting the wrong factors, ignoring differences between business models or market environments, or sizing the position badly. Being able to distinguish a data failure from a thesis, portfolio-construction, or ordinary uncertainty failure is what makes the process genuinely improvable.
Exactly. Good data improves the visibility of the system, but it does not automatically improve the decisions made from it.
A useful investment process should be able to trace each bad outcome back to the correct layer: inaccurate data, inappropriate classification, weak factor weights, poor portfolio construction, or simply unavoidable uncertainty.
Without that separation, every failure gets blamed on “the model,” and the process never learns what actually needs fixing.