Still I would argue that in typical scene where you have single stationary directional light and large landscape you should use automation to skip all components that can’t cast shadows.(see code above) For LOD bias selection human interaction may be right choice.