A lot of developers don't really need to know much math. UI, web techs, business rules etc, it's mostly about knowing the framework and the problem domain. Though, Comp Sci is basically a branch of mathematics (though CS != software engineering), so in that sense you can argue that you need to be good in it. As someone else said, you might not need to have specialized calculus/algebra knowledge, but if you have problems picking up basic concepts even with some effort, then I'd consider it as a red flag too.
And even in cases where you don't need to know math to code something, it might help a lot to make it more efficient. I'm probably better in math than in programming (I competed at one of
these, and I try to find jobs where I can use this, and I had a case like this very recently:
There is some code that caches something so it's faster for look up. This worked fine for small sizes, but scaled terribly. Imagine that you organized your L2 cache so that searching and finding an item took 3-4x more time on average than if you just fetched it from RAM directly, lol, that's very similar to what was happening, and no one noticed. The original algorithm was simple and it worked, and people aren't versed in math so that was it.
I thought of a better algorithm, and my first version looked complicated so I looked up on the internet, and found a better way to organize it, and because I was thinking about it, I could immediately understand everything, even though I haven't done anything from that domain since college (fyi, the stuff I use is an augmented version of this:
http://en.wikipedia.org/wiki/Combinatorial_number_system, you can use this to develop a way to index subsets of fixed size). The solution I proposed improved significantly on all metrics: it used about 30x less space, so instead of > 1MB array in extreme cases, it uses only about 35KB, and the lookup is faster by 3-4 orders of magnitude, up to 30,000x. Since for certain operations we do thousands of these lookups at once, the difference was quite significant.