The Bechdel Test.
If you're not familiar with the Bechdel Test, it goes something like this: A work passes the Bechdel Test if it contains at least one scene where two female characters talk about something other than a man. If you don't understand why that would be significant, this is probably not the post for you!
Now, do not get me wrong. I think it is wonderful when works pass the Bechdel test. I think it's horrifying that this is actually noteworthy, but it's still a good thing nonetheless. I have no problems with the Bechdel Test itself.
What I have a problem with is people glorifying it as some sort of litmus test for genderfail. Because it isn't. It so isn't. And that's partially that the test was never intended to be that, and part that such a thing is actually impossible.
The Bechdel Test, in its original form, was intended to be applied to a specific type of fiction: movies. And in that, it works a lot better: Movies are short. They contain only a few dozen meaningful conversations at best. A single conversation out of that is a lot more significant than in an epic novel, where people never start talking, or a serialized series that runs for years on end.
The Test works better over short distances. Longer series will pass it simply based on longevity, and when they do, it's less meaningful. Shit, even Naruto manages it, and I think we all know that Naruto is a clusterfuck of genderfail. MIRACLEHATE.
Additionally, the Test is not intended as a measure of quality: it's an entrance requirement. The character who proposes the test will not see a movie unless it passes. The question of how she would know which movies do and don't prior to the advent of the internet and blogs specifically intended for that purpose aside, taking a gatekeeper requirement like that and using it as a measuring stick for quality renders it meaningless. Applying it to things that it was never intended to apply to is like using measuring tape that only shows metric and expecting it to tell you how many feet tall you are. And you don't know the conversion rate. Shut up, it made sense in my head.
And of course the other reason why treating it as a genderfail litmus test doesn't work is that it's impossible to have a binary test for something as context-dependent as genderfail. Context is everything. You can't have any sort of meaning without it. But the Bechdel Test, and all other litmus-style tests, don't address context: they depend entirely on a very limited set of facts, and analyze based only on those, disregarding context.
The results, while handy for categorization, are not meaningful. They do not speak to the quality of the piece, either as a piece of written work, or in terms of genderfail. And that's fine, since that's not what it was intended as in the first place.