Rainer, I think you missed my point. The ability to maintain a single secret is difficult enough, the likelihood of maintaining a number of secrets with an ever increasing number of participants such as those required for stealth mode, substantially declines over time, regardless of the efforts in place to maintain it.
Of course your assumption that minimizing the number of people (n), communications, etc, minimizes the probability of unwanted leaks is mathematically factual, but in practicality it's naive. The numbers of people, communications, etc. involved in POET's stealth mode has had to increase, otherwise they could not move forward.
Regarding the summary, the key takeaway there is "All the models and countermeasures here make suppositions that are sometimes profoundly false". The study attempts to measure probabilities for something we all know, even one discretely managed secret likely has a measureable shelf life. As stealth mode involves multiple secrets and increasing numbers of participants, it's mathematical and practical potential for ongoing success is very likely to decline over time.
That time is rapidly coming to its inevitable end, whether they want it to, or not.
