I believe a balanced approach should render best ffb results while allowing some margin of user safety.
Game developers have expressed the complex challenges surrounding consistent ffb-signal output without inducing clipping. Adding to that challenge, is the use of mod-content, where even far greater inconsistencies may occur. Beta-testing of titles may also produce very extreme signal anomalies.
The short and simple fact is that such inconsistencies do exist and it’s up to us to use settings that provide the full ffb-dynamics while also operating the wheel at relatively safe power output levels.
Use of in-game or 3rd-party ffb-clipping tools can help establish a good baseline power output level and ffb-gain setting. Some titles (AC, RaceRoom, Pcars2, AMS2) offer live per-car ffb adjustment, that can be very useful for tuning and testing of ffb levels.
I doubt the difference in extreme approaches result is significant detail benefits, and certainly don’t warrant risking injuries.