There are a few points here:
1) The statistics suggestion is a complete misdirection, and this advice should be completely ignored, along with the person who gave it to you. Statistics is an interesting topic, but not at all useful in game programming (except maybe for a few esoteric approaches to esoteric topics, maybe, like, drawing clouds).
2) (Not that you seem to but...) it's not uncommon for programmers to make the mistake of assuming that they can just learn everything on the job, but most science topics (including math) can not be effectively learned this way. With these, one needs a much more structured approach, building an elaborate structure of ideas, with each new idea built on top of the previous. You could certainly program games with a few equations that you learned to use from a game programming book, but it's unlikely you'd ever have the ability to solve problems that you hadn't already seen solved somewhere else.
3) The best way to get comfortable with math is to solve lots of problems, and not on the computer, but with pencil and paper. For example, you can easily write a program to test that sin2+cos2=1, but to prove it, you need to understand it.
4) Of the topics you'll need, trig is the most time effective place to start. Geometry would be a bit useful, but probably not so much. Another useful topic is linear algebra. Calculus is also useful for calculating trajectories that have acceleration (and gravity), but it's a much bigger topic and involves so many new ideas that it's probably a bit difficult to pick up on your own. Maybe for this topic it's best to try to glean a few useful approaches and equations.
Final suggestion: I recommend starting with trig, and use a book that gives concise explanations followed by lots of problems that are solved in the back. For example, Schaum's Outline of Trig for $13, would probably be a good choice. You don't need to solve a every problem in the book, but work them until you're comfortable, and then move on.