Burland and Federico 1719 Orr and Cherubini (2003) compare different predictive equations of (i) the bearing capacity of driven and root piles and (ii) the at-rest earth pressure coefficient K0(NC) of normally consolidated clayey soils. They assess the reliability— or, quoting Orr and Cherubini (2003), “the dependability”— of the considered calculation methods in terms of the ratio Qcalc /Qmeas between calculated and measured values, and of two functions (namely: the ranking index RI and the ranking distance RD) of this ratio. The writers can see the logic of using RD in preference to RI. Orr and Cherubini (2003) introduce a somewhat arbitrary scale of conservatism in Table 1 which, in the writers’ view, is unhelpful and can be misleading. Conservatism (or unconservatism) is measured by the percentage of K values (Qcalc /Qmeas) less than 1. Consider a case in which a method gives values of K that are all less than 1 but are all greater than 0.99. By any standards such a method is both accurate and precise. Yet, according to Table 1, the degree of conservatism would be ranked as 1, and it would be defined as very unconservative (i.e., very unsafe). This is clearly not the case. Moreover the terms “conservative” or “unconservative” will depend on the context of the problem being studied and what is conservative in one problem may well be unconservative in another. We can take the predictive accuracy of the at-rest earth pressure coefficient K0(NC) as a case in point. Orr and Cherubini (2003) consider seven equations. Among these, there is the one proposed by the writers (Burland and Federico 1999) given by eq. [8] and expanded upon in Appendix A. This equation, based on the Modified Cam-Clay model, requires two input constants, i.e., the effective angle of shearing resistance φ′ and the Skempton's pore pressure coefficient at failure Af,i. Both these constants are obtainable from a single isotropically consolidated undrained compression (CIUC) triaxial test with pore pressure measurement. Orr and Cherubini (2003) state: “Equation [8] by Burland and Federico is very unconservative, tending to overpredict K0(NC)” (i.e., it is very unsafe according to Table 1). Taken at face value such a statement could be very prejudicial. Yet there could be situations in which a small overestimate of K0(NC) would lead to a conservative overestimate of loading—as in the case of a retaining wall. In summary, the writers think that the terms “conservative” and “unconservative” are unhelpful and can be positively misleading. In this case of K0(NC) the terms “overpredict” and “underpredict” would be more appropriate. Moreover the writers would not recommend the use of Table 1. Orr and Cherubini (2003) further state: “Burland and Federico (1999) anticipate, however, that this equation will be improved as laboratory testing improves.” Orr and Cherubini (2003) probably mean “...that, as laboratory testing improves, agreement with this equation will improve.” In fact, in their original paper the writers wrote “...as already put in evidence by Roscoe and Burland, this model tends to somewhat overpredicit the observed values of the coefficient of earth pressure at rest K0(NC). However, with better instrumentation and quality of testing, it is likely that higher values of K0(NC) may be obtained, thereby giving improved practical significance to the predictions of the proposed equation.” Finally, many of the columns in Table 6 are incorrectly related to the equations and a corrected version is given below (see Table C1). Note also that in Table 7, K0(NC) should be replaced by K and in the second last row RD should be replaced by K.