I am not a teacher, but I'm studying computer science (though, not some game degree).
We tend to get some people every year who think they can 'do computers' when all they know is how to use an office suite or who think they know how to program when all they ever did is copy paste together a webpage. My university works hard to keep their numbers low, but there's always some who don't take hints.
At my university, when there's an open day, faculties display some of the projects they're working on or have been working on. They usually exhibit the resulting programs in the lecture halls.
To help explain the projects, they hang up posters with images, text and mathematical formulas. This might be e.g. a poster explaining pathfinding in RTS games, which also explains A*, or a poster explaining projections, showing the math behind projection matrices, or a poster explaining traffic lights modelled using petri-nets. Often, problems are described in a way visitors can relate to (e.g. traveling salesman as the shortest sightseeing tour or taxi route), with an explanation about how the problem is solved.
This serves a twofold purpose: first, visitors can see how the topic in question works, which might spark further interest. Second, there's math right there, so it doesn't come as a suprise. (also, as a side effect, visitors informed enough will often grok it right there.)
Upside: Let's people take a look at interesting stuff and the concepts behind the 'magic' being pulled of.
Downside: Lots of preparations necessary.
Faculties also run orientation courses before the semester starts, where students can get a refresher on programming and math. During orientation, prospective students are also toured around the campus and they're being helped with finding the informations they need to put together their lecture plans. At this point, students will be shown the amount of math courses (The information (what courses need to be taken and a rough outline of their content) is also freely available on the faculty websites and in study guides, so students can see what they're up for long before they sign up). During orientation, faculty members and students helping with orientation will usually talk about their own experiences (stories about pulling all nighters and working more than a day on exercises/assignments are always being told).
Upside: People know what they're in for and have an easier start.
Downside: Needs preparation. Websites need to be kept up to date. Some prospective students skip this optional preparation.
Then, most of the math courses are put at the start and are quite grindy, persuading those who didn't get the broad hints upfront to go looking for greener pastures. Also, most of the interesting courses come later and have the basic courses as a requirement (e.g. graphics programming after algorithms and data structures and, above all, after the math course which covers vector spaces).
Excercises for the basic algortihm and data structures lectures require (after a one time tutorial) programming from the get go. This is another major factor contributing to drop outs. (Students can flunk up to four excercises in this lecture... after four weeks, thoose who can't program are gone.)
Downside: Many people quit after the first semester or change to other fields. About 50%-75% loss total after the second semester, I'd guess.
Upside: The remaining people know what they're doing.