Let's say you have a response with two levels, A or B. The predicted values for logistic regression are probabilities for the two levels. When you do a logistic regression in JMP, and save the formulas, it creates a column called "Most Likely...". This column does not use a cutoff, but simply identifies the response level with the highest predicted probability. For example, if the predicted probability for level A is 0.72, then the predicted probability for level B is 0.28. The Most Likely level is A.