An audit trail is an append-only record of actions, actors, and timestamps.

An audit trail is an append-only record of actions, actors, and timestamps. In research, it turns a claim into something that can be checked against recorded evidence rather than memory or narrative. The core idea is simple. Each material action is logged when it happens, the log is preserved, and later reviewers can reconstruct who did what and when.

For research workflows, the mechanism matters as much as the definition. Sonar describes a research-to-publishing process in which work moves from research records toward publication with preserved provenance and reviewable history. In that framing, an audit trail supports accountability by tying outputs back to the sequence of recorded steps that produced them. This is what makes a statement about a dataset, parameter choice, test run, or publication decision open to verification instead of assertion alone.

An append-only design strengthens that process because new events are added as new records rather than replacing prior ones. If a parameter is changed, the meaningful question for review is not only the final value but also when the change occurred, who made it, and what earlier state it superseded. A durable event history lets a reviewer examine that progression directly.

The same logic appears in Sonar’s treatment of backtest overfitting auditability. The backtest overfitting audit tool is presented as an audit-oriented control for research activity, with emphasis on documenting the path from experimentation to reported result. That kind of record helps separate a research conclusion from the sequence of trials that preceded it. In practice, an audit trail is useful when it preserves the chronology of runs, changes, and decisions so a reviewer can inspect the process that generated a reported metric or conclusion.

This is especially relevant when a research claim depends on methodological choices. Sonar’s glossary entry on the deflated Sharpe ratio explains that evaluation of a strategy statistic must account for the number of trials and the risk of selection effects. That makes the surrounding research history important. If many variants were tested before one result was highlighted, the evidentiary value of the final figure depends in part on whether that trial history is recorded. An audit trail provides the record needed to connect the reported number to the research path that produced it.

An append-only audit trail helps make research claims verifiable by preserving a time‑ordered record of actions and decisions.

Covered in depth in the From research to publishing pillar hub.