The word gets used to mean disciplined, or quantitative, or unemotional. It means something narrower and more useful than any of those.
That is the whole definition. What is held, when it changes, and what triggers the change are specified in advance and applied when the trigger occurs.
Not "we have a disciplined process." A specified rule, written before the circumstance it governs.
Consistency across accounts and across time. A documented reason for every decision that existed before the decision. And immunity to the specific failure where a process quietly changes under pressure and nobody acknowledges it did.
The last one is the real prize, and it is not visible until it matters.
The rule cannot respond to what it did not anticipate. Presented with a genuinely novel situation, a systematic process will do the wrong thing, confidently, because following the rule is what it is for.
Anyone describing systematic investing without saying that is describing something else.
A rule that has never been wrong has usually not been tested. It has been fitted.
Rules found by searching history for what would have worked will always look excellent on that history. Whether they describe anything durable is a completely separate question.
The more parameters a rule has and the more variants were tried, the more likely the result is an artefact of the search rather than a property of markets.
Few parameters. A reason the rule should work that is not "it did." Testing on data the rule was not developed on. And publishing what the approach is bad at.
It is not a guarantee, a prediction, or a claim about outcomes. It is a description of how decisions get made. Nothing here says a systematic approach performs better than any other, and that is not a claim I will make on a website.
Read next
Systematic investing, defined →