A historical result is useful because it makes an idea testable. Its value depends on the quality of the question, the assumptions and the data behind it.
Start with the decision rule
Before inspecting a return curve, describe what the system is trying to do. Which market behavior motivates the rule? What evidence would challenge that explanation? A clear hypothesis makes it easier to distinguish a coherent process from a setting that happened to fit one sample.
Make the experiment legible
Time period, instrument universe, data treatment and cost assumptions shape the outcome. Results become easier to compare when those conditions are documented with the system version. A single headline number cannot communicate that context.
Examine where the behavior changes
Different periods can reveal different weaknesses. Trending markets, extended consolidations and abrupt transitions may place very different demands on the same rule. The point is to understand those differences, including the parts that do not support the original idea.
Carry the limits into the next stage
Historical testing cannot reproduce every detail of live execution. It can guide the next experiment: what to observe, which assumptions to challenge and how to compare expected behavior with actual behavior. That is a stronger starting point for engineering.
The useful output of a backtest is a better understanding of the system, including the reasons it may fail.