Put another way: as far as natural selection goes, women undergo a selection process that, depending on their fitness (and chance, of course), replaces them with maybe 0-6 copies of their genes; men, on the other hand, can be replaced with many more (even if, on average, they are not, the distribution is much wider), and those that hit the "jackpot" will tend to very quickly spread their genes throughout the pool.
The extent to which this actually happened is unclear; I don't know off the top of my head if there are good ways to investigate the landscape of possible fitness outcomes based on only the genome, but if this did happen to any significant degree we'd likely find that wild genetic variations stabilize more quickly on the Y chromosome than on the X (we'd also find that wild variations elsewhere, good and bad, tend to spawn more often from males rather than females, but I'm reasonably sure there's no way to see that without a full DNA history because any genetic material not on the Y chromosome ends up in females as well as males).
I'm currently very interested in this because I think it might be a useful approach in genetic algorithms to cut a balance between the two common strategies for selection - the male replacement strategy (weighted by fitness) is very useful for spreading good characteristics quickly through a population, but the female one (flat selection via a fitness cutoff) is good at preventing one particular trait from forcing fixation of all the genetic cruft that happens to sit beside it.
For instance, imagine that we're evolving zubbins, and each one has two traits, color and size. It turns out that the optimal zubbin for this environment is red and tiny, where "red" is extremely important, and "tiny" less so. If we started out with a population mixed between tiny green zubbins and large red zubbins, a fitness-weighted reproduction strategy would cause the large red zubbins to very quickly overtake the entire population, and we'd lose the "tiny" trait altogether, and the only way it could be recreated would be by mutation (which is a relatively weak evolutionary force, typically). On the other hand, if half the population was selected by a cutoff, there are more "spots" available for less fit zubbins, some of which will be tiny, so eventually we'll see some tiny red zubbins, which is what we're after, and these will still reach fixation quickly enough because of the massive benefit that the male offspring will have.
If only the male reproductive strategy was in play, the crappy "large" trait hitchhikes along with the useful "red" one, destroying the "tiny" trait altogether; conversely, if only the female strategy was happening, fixation would take much longer.
What I'm not sure about, though, is whether there's any real benefit to separating the two strategies (i.e. creating both male and female classes) rather than coming up with one method that cuts a good balance between them (offering a slight extra benefit to highly fit creatures, but just setting a cutoff at some point). There are a lot of other reasons that sexual reproduction is evolutionarily useful in the real world, and many of those don't carry over to code.
As a rule, I definitely think that playing with the rules of evolution, and more specifically, letting those rules (perhaps implicitly) evolve themselves, will be a key idea if we want to exploit evolution for all it's worth; finding a way to let the process significantly improve itself is, IMO, the Holy Grail of evolutionary computing. Evolution is, in some sense, a process of emergence that's been optimized by emergence, and I suspect that our best shot at creating intelligence may be to follow that pattern and attempt to achieve an evolutionary process that's optimized by evolution.
What results when you have intelligence optimized by intelligence is another matter, and (IMO) a far more dangerous one, but that's another story for another day...
Re: writing something up, I'll add it to my todo list, hopefully it will happen eventually, though I'm fairly slow at more formal writing... I am planning to do a series of posts on evolutionary methods in programming, where I want to cover (and hopefully prove useful) a lot of biologically inspired things usually left out of vanilla genetic algorithms/programming (speciation, gender, chromosome/gene separations, automatic evolution of coding schemes, embeddings of meta-evolutionary abilities within the normal evolution process, and hopefully more), but I'm lagging on actually trying out all the things I'd like to, and I also want to find some more compelling practical uses for this stuff...can't make any promises, this is currently a total side-project for me, but I'm plugging away at it, so I'll do my best to write it up eventually. I'll start thinking more about what might be possible to infer from real DNA; my brother is working on his bio-informatics PhD and I've had a lot of conversations with him about this type of thing, so I know that some of the stuff they're able to pull out of the data is truly incredible, they very well may have the data necessary to answer questions about fixation speeds in each chromosome, which would be extremely interesting whatever the results were.
Sent you an e-mail, and added info to my profile (didn't realize that it wasn't already available from the e-mail field, that should be more clear IMO) - if you really think any of this is worth pursuing, by all means let me know.