Additive transitivity defined by S.Alonso et all in "A consistency based procedure to estimate missing pairwise values" page 8 is apparently similar to the definition provided by Z.Wang in "Goal programming approaches to deriving interval weights based on interval fuzzy preference relations" page 7. 

But the example in Wang's paper page 21 does not satisfy the definition given by Alonso. 

Is there a problem in my understanding of the definition. 

Thank you.