A few years back I followed a distance master degree in maths at UPMC (Université Pierre et Marie Curie, in Paris). At the time the university sent course documents and exercices by post every week, solutions to exercices had to be sent back, and students had to be physically present for the final exams. Subscription was cheap (a few hundreds euros I think). I don't know to what extent all this has changed since then; see here for the Bachelor degree: http://www.telesciences6.upmc.fr/fr/licencemath.html
I did not get my master degree in the end. Reasons are:
I was already in the process of getting a master degree in engineering, so pressure was low to get another master degree. And I think a bit of self-pressure is required with any distance solution, because no one but you is there to push you on a daily basis.
Not only pressure but also emulation. Being part of the same boat with other students is important to discuss ideas, problems, wrong solutions, etc. and sometimes to do something else. I was lacking interactions with peers.
Some courses such as stochastic processes were entirely new and strange to me, and I had much difficulties to understand them without a teacher. At the same time the temptation was high to spend more time on courses I liked such as groups and representation theory. It means that a lot of self-discipline is required, more than for regular courses where the agenda "forces" you to tackle with all topics.
For maths you need to do exercises to make progress, and you need to experience it to know how much it is true. The few exercises that were asked were maybe not enough, and I did not have the motivation to make additional exercise sessions by myself. The danger is that reading maths (i.e. reading the printed document courses) is something very pleasant I think, but this is not very rewarding if not completed with taking notes, doing exercises, asking questions, etc.
I guess the development of MOOCs nowadays can moderate several of the points mentioned above; I have followed a few MOOC math courses recently and I have been quite enthusiastic about them - but this is certainly another topic by itself.
Some years later I started a PhD in parallel to my regular professional activity. I worked on my PhD during evenings, week-ends, and vacations. I did it for me, not for carrier purposes. Concretely I was rarely in the lab, I used to see and email my advisor and a few other researchers from time to time, I attended a few conferences only, so this can probably be called a distance PhD degree. I succeeded in this enterprise, but I think the context was fundamentally different from that of the master degree:
this was research, not education, so I did not really need to learn things by heart or to force myself to work on topics I disliked; I naturally followed the path of maths I felt comfortable with, and this is where I found results: magic!
research is not a linear process, so I think your discoveries depend more on a maturing process than on the total time you spend working on it.
Hope this can help.