The line between factor and noise
When outcomes resist neat explanation
Factor decomposition is often sold as a total solution. We know the feeling when results
don’t add up, even after careful analysis. We help clients get comfortable with those
gaps, focusing on what can truly be measured and communicating what is left unexplained.
If a recent report or model output leaves you with questions, we are here to discuss
it—no pretence, just honest conversation.
Ask a question
Facing the unknown: idiosyncratic risk in returns
Factor analysis promises answers. Idiosyncratic effects remind us that noise always remains. Our practice distinguishes these lines with care, sharing both data and its limits.
Where measurement ends and unexplained risk begins
- Splitting factors from idiosyncratic effects: See how we isolate measurable factor contributions and spot when remaining return is unpredictable.
- Diagnosing unexplained returns: Understand the practical and statistical signs that a return component cannot be reliably attributed to any known factor.
- Distinguishing random patterns: Recognise when an apparent pattern is simply noise, using published research and field examples for context.
- Communicating uncertainty with clarity: Discuss how to communicate these limits to stakeholders, helping avoid false certainty and encouraging honest reporting.
Our framework priorities
Get comfortable with the reality that some return components will defy explanation, even with sophisticated models. Learn what can be measured, and where humility is required.
What keeps our analysis realistic and honest
We emphasise candid communication—naming where even the best models hit a wall. Our process uses lived examples to keep the conversation grounded, never just theoretical.
Inside our review: boundaries and blind spots
Spotting the unexplained in charts
A team member points out sections of a return chart where the model cannot explain
outcomes, sparking discussion about possible sources of idiosyncratic risk.
Where models diverge
Analysts review a case where different models disagree, using it to illustrate the boundaries between factor-driven and idiosyncratic returns.
Marking the boundary in analysis
The team debates where idiosyncratic risk takes over, marking the transition point
between measurable and unexplained outcomes for a portfolio.
The benefits of naming what models can't capture
No method explains everything. We support honest, open reporting that names boundaries, not just successes.
Making uncertainty a visible part of analysis
We call out the line between what’s knowable and what is not, helping clients see the actual scope of factor models. Our practice resists the urge to explain away unexplained returns, keeping risk visible.
Noise named as noise
Limits are visible
Transparency over false certainty
Stakeholders deserve honest reporting. We explain findings in plain language and don’t gloss over what remains unexplained. This openness helps avoid costly misinterpretations.
Communication is open
No overconfident claims
How we keep uncertainty visible and useful
Every unexplained outcome is named—not hidden. Our process avoids overconfident claims, supporting informed decisions at every step.
Tools and habits for honest decomposition
There’s pressure to explain every outcome. Our process values the honesty of uncertainty, helping clients avoid stories that are too neat to be true.
Separating what is explained from what is not
We provide tools for splitting observed returns into explained and
unexplained components, but never force a story when the data resists.
Connecting theory with lived experience
Our team studies both theoretical research and real case histories, showing how idiosyncratic risk appears in daily analysis.
Naming overfitting and reporting limits
We teach teams to spot overfitting and over-explanation, using transparency
as the core value in reporting to decision-makers.
Humility in the face of noise
We value humility. Sometimes, the smartest answer is admitting uncertainty,
not inventing an explanation.