This paper introduces a new MILP modeling to find impossible differential (ID) distinguishers and attacks. Standard models for ID are negative models, in the sense that a differential is impossible if and only if the model has no solution. Our new modelling technique focuses on probable ID, differentials that are probably impossible. While this might lead to false positives, the main advantage is that searching for such probable ID can be achieved through a positive model. This facilitates the search for the best impossible differential attacks without first exhausting all possible ID distinguishers on a target. We also propose to simplify the modelling by only considering two possible states for internal cells: inactive and unknown. In this case there are no longer direct contradictions but only indirect ones, assuming that it is impossible that all cells are inactive. With these two simple ideas, we are able to retrieve the longest impossible differentials distinguishers on MIDORI, SKINNY, PRESENT, SIMON, Simeck and SPECK. Furthermore, as the model looking for candidates is based on satisfiability, it can be incorporated in a larger model which looks directly for the best attacks in order to enumerate the distinguishers in the order of the complexity of the associated attacks, which we did for the AES, ARADI, SIMON and SKINNY.