Historical tests ask a conditional question
A backtest applies a rule to past conditions. It asks what a specified process might have produced under stated assumptions; it does not document actual investment activity. The Investor.gov performance bulletin makes this distinction explicit. In our proposed records, a historical test would always carry a hypothetical label and a description of its data and assumptions.
We would retain the period selected, the rule version, the comparison and the point at which information entered the test. Any change to the sample or implementation would be recorded. A historical result would not become more reliable merely because the presentation is precise or the chart is attractive.
Forward simulations test a process in sequence
For our planned work, a forward simulation means applying a previously documented rule as new information arrives, while keeping the activity simulated. We would record proposed decisions at the time they are made and distinguish assumed execution from observable market information. This can create a reviewable decision sequence without implying capital was committed.
The simulation protocol would define when inputs are captured, how unavailable prices are handled and what happens if a rule cannot be evaluated. Changes would start a new version. A simulated fill remains an assumption, even when based on a quoted price; it should never be presented as a completed transaction.
Live records require a different evidence trail
A live result would need records of actual transactions, cash movements, valuations and applicable expenses. We would reconcile these records before describing an investment outcome. The label “live” would not substitute for evidence about what was held, when it was held or how its value was calculated.
No live Arccerno investment results are presented on this website. Likewise, the research agenda is a plan, not a completed simulation. Keeping these states separate is part of the content standard we intend to maintain as the research process develops.
Compare like with like, then explain differences
Our proposed review format would disclose currency, measurement period and the treatment of distributions, financing, trading costs and other relevant expenses. Missing cost information would be identified, not silently treated as zero. A comparison would use a stated basis for both the research rule and its benchmark.
Reviews would separate changes in the research view from deviations in execution or measurement. We would keep unfavourable periods and unresolved discrepancies in the record and explain the next investigation required. None of these categories supplies a promise about future outcomes. The aim is an honest account of what a result can—and cannot—support.
THE ESSENTIAL POINT
A result should carry its evidence type, assumptions and limitations wherever it is presented.
Sources & context
Supports the distinction between hypothetical and actual performance, and the need to understand fees, assumptions and comparisons.