I’d like to propose a feature that would bring SimBrief closer to real-world airline dispatch systems: an optional feedback loop that learns from actual flight results and applies a fuel/performance bias over time.
Simbrief already does a great job with airframe performance profiles, weather based planning, and statistical contingency models such as B343 with ACF90/ACF99. However, unlike real airline operations, it does not “learn” from completed flights.
In the sim, fuel accuracy can vary based on the simulator, aircraft add ons, and pilot technique (CI usage, climb profile, etc. Even if simbrief is very close, users may see consistent deviations in one direction. Right now the only way to tweak that is by using the static “Fuel Factor” and editing fuel manually.
Proposed feature:
Post flight feedback input that allows users to log planned vs actual trip fuel, and/or block fuel vs landing fuel. An optional addition would be a place to indicate if there was a reroute, holding, etc. This could be accomplished via a manual entry or API integration with established flight tracking tools such as Stratos, SmartCars3, or Volanta.
Simbrief would take the data and calculate a rolling correction for the airframe and apply it to the trip fuel prediction and enhance the ACF90/99. For example, if a user consistently landed very close to plan, retain or slightly reduce contingency and if variability is higher, increase contingency. This would reflect actual user flying rather than a generic model.
Suggested user controls: On/Off toggle, Reset History, Ignore Outliers, Show current bias in OFP.