What belongs in a performance disclosure

A performance disclosure should state the assumptions behind the result, the cost treatment, the sample limits, and whether the result is hypothetical. Sonar Sciences’ guidance says this information should be placed adjacent to the relevant chart, table, or metric so the reader can interpret the figure with its limits and conditions in view.

Guidance: What belongs in a performance disclosure (Sonar Sciences)

A performance disclosure explains the conditions under which reported results were produced, what the results leave out, and why they should not be read as realized outcomes. For quantitative research, the disclosure belongs where the results are shown so the reader can interpret a chart or table with the relevant limits in view.

Sonar Sciences’ research-to-publishing guidance says a report should disclose assumptions, fees, sample limits, and whether a result is hypothetical. It also says to place these disclosures adjacent to the corresponding chart, table, or metric rather than burying them elsewhere. A return series, Sharpe ratio, drawdown figure, or other summary can change meaning materially when execution assumptions, cost treatment, sample scope, and the simulated nature of the test are made explicit.

In practice, that means identifying the modeling choices that govern how signals become trades and how trades become reported statistics. If sizing, rebalance timing, fill conventions, universe selection, signal delay, or data handling choices affect the result, those are part of the assumptions the disclosure should name. Sonar’s publishing guidance frames assumptions as a required component of any performance presentation because they define what the reported number actually measures.

Costs should also be listed directly with the result. Sonar’s guidance specifically calls for disclosure of fees, and its overfitting audit tool reinforces why this matters by separating a gross view from a net view after estimated fees and slippage. A strategy can look different before and after implementation frictions are applied, so the disclosure should tell the reader which cost components were included and whether the figure is gross or net of those costs. If a table reports net results, the cost basis used to arrive there should be stated alongside the table.

Sample limits are another required part of the disclosure. Sonar’s research-to-publishing guide calls for sample limits to be disclosed, and its materials on overfitting and statistical interpretation show why. The backtest overfitting audit highlights the risk of selecting a model from many trials and reports diagnostics that depend on the number of tested variants and the structure of in-sample and out-of-sample evaluation. The glossary entry on the Deflated Sharpe Ratio likewise explains that statistical interpretation depends on the number of trials and the likelihood that an apparently strong metric could arise after multiple testing. For a performance disclosure, the practical implication is that the reader needs to know the sample boundaries and constraints that qualify the result, such as test period coverage, number of observations, and any other limits that affect how much weight the metric can bear.

The disclosure must also state clearly that the results are hypothetical when they come from backtests, simulations, or model-based reconstructions rather than realized trading records. Sonar’s publishing guidance identifies hypothetical status as something that should be disclosed with the result. This matters because hypothetical results are produced under assumptions and do not reflect all of the uncertainties and frictions of live execution. Stating that the results are hypothetical tells the reader what kind of evidence is being presented before they interpret the figure.

It is part of the communication standard. The research-to-publishing guide says disclosures should sit adjacent to the chart, table, or metric they qualify. This placement improves clarity because the relevant assumptions and limits are visible at the point of use. The reader does not need to search for footnotes in another section to determine whether a result includes costs, how broad the sample is, or whether the figure is simulated.

A complete performance disclosure therefore does four jobs at once. It identifies the assumptions behind the reported result. It states the cost treatment, including whether fees and slippage are included. It describes the sample limits that constrain interpretation. And it says explicitly when the result is hypothetical.

Drafted with AI assistance from cited sources. Reviewed and approved by Sonar Sciences Quant & Research Team.