Discussion. Concrete Cube Strengths - What Use Are Statistics?

Abstract

With regard to &19-21 of the Paper, I think that the Author has been over-zealous in his attack on the characteristic strength definitions. Without losing oneself in semantics, the definitions may become more palatable to the Author if the word 'expected' is treated as a technical term which denotes that one is making a statistical statement (i.e. a statement about a population of results that is exactly true in theory only). 55.With regard to the CP 110 compliance criteria and the possibility of a 'logical impasse' ( §30), I think that the Author has hit on a very real practical disadvantage-when it occurs-but according to the statistical philosophy of concrete strength the probability of occurrence of the 'impasse' is small (but non-zero). In other words, the philosophy accommodates an 'impasse' but predicts that it should not happen very often; hence on these grounds it is not deemed important (rightly or wrongly!).56. To put the 'logical impasse' in greater statistical perspective (a) each individual result-x-is required to be such that x > 0.85 k (where k denotes the characteristic strength). Since, by definition, only 5% of results are expected to be less than k, the probability of x 0.85 k (i.e.area A, in Fig. 2) is less than 0.05. In fact it is given by 0.

The study of the conformity criterion for compressive strength of concrete based on order statistics

“…W przypadku produkcji pocz tkowej, obowi zuj ce kryteria wydaj si by niedostatecznie uzasadnione [3,4] a przyj te wspó czynniki oceniane s krytycznie przez wielu autorów [np. 5,6].…”