Independent backtest audit for funds and signals

The problem

The selected curve does not show how many model versions were discarded.
A presentation can mix gross, net and simulated results without explaining what each part represents.
The reader needs to distinguish calculations on the file from assumptions supplied by its author.

A dated series of period returns at a consistent frequency, or the portfolio value over time, as CSV or Excel.
Identify whether the series is gross or net and which costs it includes. Keep both versions separately and explain methodology changes and external flows.
A dated benchmark CSV under Advanced options, and documented research attempts to put the selection in context.

Evidence and selection. The report distinguishes Measured, Declared and NOT Measured; the deflated Sharpe also depends on the supplied number of research attempts.
Behaviour within the history. It examines uncertainty, concentration in the best periods and differences between the recent period and the rest, as the supplied data allow.
Comparison and costs. With a compatible benchmark, it compares matching periods. Costs need their own evidence: calling a series net remains the provider's declaration.

Rigor analyses supplied files; it does not reconstruct the signal or run its code.
It does not connect to the broker, establish asset ownership or replace legal or operational review.
It recommends no investments and promises no future results. Without the file it cannot reach conclusions about that record.

Does it work for a model portfolio without trades?

A dated series supports a review of its returns. Without trades and itemised costs, some execution questions remain unmeasured.

What should we clarify before supplying the file?

State the frequency, currency, fee treatment, benchmark and model change dates. Contact Rigor to define the material under review.