In general ADS / ADM model count is dependent on the outcome of a offer. If the outcome of a offer is positive then positive count will be incremented and if the outcome is negative then its count will be incremented.
I believe you posted the exact same question on the Mesh? For people having access, this is the link: https://mesh.pega.com/message/290030. For people w/o access, the short answer is no. The slightly longer answer is no, and it doesnt matter when you use the propensity in a relative sense (like for ranking, which is what it is used for most of the time). If you really care about the absolute propensity value, the returned propensity can be transformed into the 'correct' value.