I disagree slightly with your categorization of the problem. I think your characterization, while accurate in itself, is missing the "otherizing" and "in-grouping" that happens if we only foster no-boys-allowed special interest groups.
Lets say a given women in tech event is phenomenally successful, that means that men are now precluded from attending a key event. Whether consciously or unconsciously men understand this and have a discentive to making the group phenomenally successful. The more successful it is, the worse the world becomes for the average man's career.
When it is x%/y% it puts things into a different category. In x%/y% all success is shared.
1) Men interested in attending these female focused events would have a vested interest in increasing both the size of the female event (-> more male tickets) and increasing the % of women at other events (-> more male tickets). If I, as a man, wanted to attend the YC female founders conference as it was just held, there is no way I can - so it isn't worth me spending any time on. If however, the event had an x%/y% structure, one of the ways I could gain admission would be to actively bring women into tech (I assume this would be a large part of the selection criteria by which the small number of men get admitted). Everyone's interests are now aligned.
2) The connections formed between the small number of men and the large number of women at x%/y% female events are connections that will act as a bridge to getting more women at the regular events. It's well demonstrated in organizational behavior that even small amounts of conscious cross department interaction leads to greater sharing of ideas and follow on connections in both groups.
2b) I'm concerned that most female focused groups are only creating in-group connections. Meaning the approach they're taking leads to B0->B1 and eventually B1->BN (where the existing tech scene is already at AN as a boy's club). Which is to say, they're creating a silo of connections. We shouldn't be looking at the problem as "How do we create the best shadow tech scene for women." But rather, "how do we make the tech scene gender neutral."
3) By making these groups x%/y% you put men in the tech scene in a more acutely aware state. "You can't go in there, it's women only" is easy to parse as a group of 'others.' Human psyches are pretty good at dealing with blanked exclusions -- its a rival tribe, which is fine, I have my own tribe. The situation of "You could go in there, but it's very hard to get in because they only let in x% men" makes the problem tangible, evaluative, and solvable. Long story short on this point, I think with x% men, men feel the exclusion in a stronger and more productive way. Men would feel excluded in a way that is more productive for society if events were x%/y%. In the current state of affairs, nobody can write a blog post titled "How I hacked my way into a men's ticket to the YC women's conference."
3b) Right now 90% of the tech scene has no skin in the game for getting in more women. Some may have philosophical beliefs or take a very long term view and thus work towards female events happening anyways. But, x%/y% would mean we can a much larger portion of the 90% working on the problem in a way that has tangible near term benefits for everyone.
Finally, we as a society only really permit discriminatory subgroups like this when the size of the subgroup is very very small. "Men in teaching," "Women in tech." By setting up "women in tech" groups as permitted discriminatory subgroups we are betting against these groups ever achieving parity. Because, when they do achieve anything even approaching parity they'll have to reconstitute. Think League of Women voters -- they're now a league of concerned voters of all genders, but their historical permitted discriminatory subgroup focus hampers their activities to this day. x%/y% means the group will never have to go through an identity crisis or a period where it is in the grey area of whether it is still acceptable to keep their historical membership barriers.