I don't think I'm following that line of reasoning.
Take mutations: gene transcription happens millions and millions of times, sometimes a gene gets transcribed improperly and you make something different. If the gene leads to increased survival, then it gets passed on.
Mate selection is a further extension of this. While it might have a whole host of factors that don't seem random, they're based on the person's genes and partner suitability, which was ultimetly determined randomly.
If we could actually see the tools we're using to select partners - hormone smells, face symmetry, etc. - I'd maybe consider it was random, but the fact that we're blind to almost all of it seems to suggest otherwise. On top of that, if it wasn't random, we'd have a whole lot fewer bad genes wandering around making life miserable. But since we can't see, and we let disease rule us, I'd say there's little that shows its not completely random.
Random: lack of pattern or predictability in events.
We do know why mutations happen though. It isn't as if there is just some cosmic random number generator that determines if a gene will replicate properly or not. If it fails, it fails for a reason. Lets call it, for simplicity's sake, fatigue, or even effect of some radiation (i.e. cosmic rays), or compositional weakness. In any case, there is predictability, if we could enumerate every possible cause. To quote Spinoza "Nothing in Nature is random. … A thing appears random only through the incompleteness of our knowledge."
The question of, say, mate selection sure appears random, because there is a huge range of possible outcomes if we select two people "randomly" out of the entire population of the earth (so, lets say 3 billion males combined with 3 billion females, just for an examples sake yields something like 4.4999999985E+18 possible combinations). Yet, of course, there is only a far, far smaller actual set of real combinations that could even really happen, because events don't just happen, they are caused by something. So, someone in, say, having never left rural Africa has almost essentially zero (or functionally zero) chance to mate with someone in, say, the dense jungle of South America because they will never, ever, even possibly interact, unless they specially left the area, which would then change the entire scope of the calculation and in doing so, prove that there is some element of determinism based on causes having effects.
You could only possibly mate with someone you actually encounter. The people you encounter are not random, they are wherever they are for some reason or other. And you interact with them for some reason or other. You decide to mate with them for some reason or other. I really don't understand how this is random, even if it is vastly complicated and incredibly hard to predict. We simulate it with something that approaches randomness, because that is the best we can do, but that doesn't make it actually random.
Same with a computer generating "random numbers." Even something like "GetTickCount" yields something that appears random, but really isn't. So, there are very expensive "random number tables" one could buy, that use vastly complicated data sets (like captured cosmic radiation) but even those are determined by something, and are not truly "random."