Systematic versus discretionary: a framework

Every investment decision is made either by a rule or by a person. A systematic strategy encodes its decisions into a defined, repeatable process: signals are generated from data, positions are sized by pre-set rules, and execution follows the model. A discretionary strategy places human judgement at the centre: an investor interprets information and context and decides accordingly.

The distinction is about method, not merit. Both can be rigorous or sloppy; both can succeed or fail. The useful question is not which is better in the abstract, but which is better suited to a given problem — and whether it is applied with genuine discipline.

Systematic approaches turn an investment idea into a process that can be applied consistently and tested against history. Their strengths are structural.

Discretionary judgement is strongest exactly where models are weakest: in novel situations, in qualitative assessment, and where reliable data simply does not exist. A skilled investor can weigh context that no rule anticipated.

A discretionary manager is assessed largely by talking to the person making the decisions. A systematic one cannot be, and the substitute is not more conversation — it is evidence about how the rule was arrived at.

The central question is whether the effect is genuine or fitted. Any rule can be made to look excellent on the history it was built from, and the more parameters it has the more certainly it can. What distinguishes research from curve-fitting is what the researcher did to try to break the result: testing it on data held back from the search, on other markets and other periods, and reporting the versions that did not work rather than only the one that did.

Three failure modes account for most of the difference between a backtest and a live result. Look-ahead bias, where the rule quietly uses information that was not available at the time. Survivorship, where the instruments that failed are missing from the history. And selection, where a rule is chosen after seeing which of many performed best — the outcome of a search presented as the outcome of a hypothesis.

None of that requires seeing the signal itself. A provider can decline to disclose the rule and still evidence the process that produced it, and a provider unwilling to evidence the process is making a different statement from one protecting intellectual property.

Rather than a fixed preference, the choice benefits from a few structured questions. The answers tend to point clearly toward one method, the other, or a deliberate combination.

Most institutional exposure to systematic strategies is exposure to somebody else's research. That changes the subject of diligence: the question is no longer only whether the approach is sound, but whether the arrangement around it is.

Two things matter most and neither is about the model. The first is change control — a systematic strategy is not a fixed object, and the terms on which the provider may alter it, and what happens when it does, determine whether the thing being held next year is the thing that was assessed this year. The second is dependency: what the allocator is entitled to if the provider stops operating, and whether the strategy can be run, wound down or transferred without it.

Block Asset Management's own arrangement is disclosed rather than described in the abstract, because a firm that asks managers about their conflicts should publish its own. The signals and systems behind both of its systematic capabilities are produced by an external provider founded by a partner of Block Asset Management — a related party, not an arm's-length third party. BAM does not develop those models. The partner concerned sits on no investment committee and takes no part in the decision to use the provider or to deploy a strategy that relies on it; those decisions are taken by the investment committee of the relevant structure, and a material change made by the provider returns to that committee as a new decision. The arrangement is set out in full on the governance page.

In practice, the most robust approach is rarely dogmatic. Systematic execution can be paired with human governance; disciplined judgement can be reserved for the questions models cannot answer. The aim is to capture the consistency of rules and the adaptability of judgement while guarding against the weaknesses of each.

This is the philosophy behind how Block Asset Management works: systematic where the edge is codifiable and repeatable, discretionary where judgement adds most — and disciplined risk management across both.