In case of the yaw moment coefficient N'rT the least squares fitting according to Clarke and Horn gave the following result

　

　

 The coefficients obtained differ significantly from those given by Clarke and Horn, Eq. (9). An additional ordinary regression analysis without presetted parameters produced

　

　

 The belonging scatter plots for Eqs. (22) and (23) are presented in Figs. 13 and 14. It can be seen that the approach of Eq. (23) obviously reduces the standard deviation σ.

　

　

　

 The estimation of the four damping coefficients is no goal in itself but has to lead to a definite assessment.

　

 First step is the determination of the well-known stability lever

　

　

 which in the case of a dynamically stable ship is larger then 0. However, this particular stability index ultimately gives no quantitative measure.

　

 Much more favourable is the slope of the so-called spiral curve rc=f(δR) at the origin (δR=0), which in accordance with Fig. 1 is defined as follows:

　

　

 In case of positive yaw stability the slope is negative!

　

 For an estimation of the slope characteristic the rudder force coefficients Y'δ and Nδ are needed, whereby it holds N'δ ≈ -0.5・Y'δ. In a pragmatic approach two solutions are conceivable. First, to take appropriate data for similar vessels from a manoeuvring data base. Second, to take advantage of empirical solutions like the one already proposed by Clarke (1970):

　

　

with a lift characteristic

　

　

and a real propeller slip ratio

　

　

 Λ is the effective aspect ratio of the rudder. Similar approaches are known from several other sources, e.g. Crane et al.(1989).