Just to clear up the math thing a little. I have here no way to use mathematical formulas, so lets try this in verbal from.
In a hotel we find at any given time a lot of people. And the hotel itself establishes the environment in which those people move. The basic thing is, they move in patterns which change over time. Now, that sounds abstract, but it is in fact easy. Take the mornings:
A lot of people on the floors 4-6 get up. Not all at the same time of course. We have early birds and long sleepers and everything in between. That's why hotesl offer usually breakfast from an early time to a late time. Like 7am to 10am (that's just an example, can be 9:30 for the Cecil, but I don't know their breakfast times there).
So, while we can't say an exact time for any given person, we know, the majority of guests for example from the fifth floor will move in this time down to the fourth, where they get breakfast. In mathematical terms, they establish a traffic stream, starting at 7am, when the fir come for food and ending at 10 am when the last long sleepers arrive. During this time, there is a higher probability that anyone, we meet, is on the way to have breakfast than for example, the probability he goes for the laundry room.
Now, there are some limitations in this model. For example people who want to come from the fifth have two obvious choices: They can use the stairs or the elevator. But they can't just teleport or beam or anything. Their choice of possible ways is restricted by the possible means of travel, in this case, they either have to go to the elevator or the stairs. Which essentially means, our traffic stream builds up like a stream or river in nature too. People come out of their rooms, enter some creeks (the corridor on the floors to the stairways or the elevators) and unite in a big stream that arrives in the breakfast room.
So, anybody, who travels on such a major temporary stream has technically the same probability to be noticed by anyone else in such a stream, right? Well, not entirely true. Alone traveling females for example are noticed with a slightly higher probability, at least by the alone traveling male members of that stream. Sorry, that's how we are wired.
Now, mathematical spoken, the probability to be noticed is higher for people who travel in such a stream than outside, simply because inside this stream the number of potential observers is higher. And the probability to be noticed in such a stream grows and grows, the more a certain person shows any kind of unusual behavior. Whereas "unusual" depends on the social consensus of the environment. Loud singing would cause more attention in a hotel elevator than in a chorus ... well, that depends also on how good you sing. But it makes the problem a bit clearer.
The beauty of this model is, we can actually determine for every time at least a first impression, where are the zones with low traffic and where are the zones with high traffic at a given time. Zones with higher probabilities to be noticed and zones with lower probabilities.
I hope, that makes the basic idea a little bit more transparent. I left a lot out, but at least, it looks without equations only half as scary.