Therefore, at present it is better to continue with your way of fixing the trailing SL. Meanwhile as per your suggestion I will keep track of the SL with this Lin Reg system also. When something useful comes out I will put it here for scrutiny.
BTW be careful with curve fitting. Either mentally or using automatic back testing methods. The parameter values that you select will show you best result in hindsight but might fail miserably in real life.
Be skeptic very skeptic