Now, as I've shown above, some conspiracy theories are in time revealed to be fact, so the Trump-Russia connection might very well be true, but at this point, from what the public has actually seen, it's still a theory.
In scientific usage (and I know we're not talking science here, but I'm offering an analogy, so bear with me), the term "theory" has nothing to do with whether a model has been confirmed or not. A theory is a model that codifies and explains a set of data ("facts") and makes predictions that can be tested to determine its reliability. A theory that's been supported by the evidence is still called a theory, for instance relativity theory or quantum theory (or music theory or color theory or game theory -- these are not assertions that a thing exists, but rather models that codify how and why it functions). A theory that's contradicted by new evidence is modified or abandoned in favor of a theory that fits the evidence. "Theory" doesn't mean true or false, it means
testable.
And that's the thing about conspiracy theories. Of course there are
some conspiracies, and if we approach the evidence rationally, it will confirm some conspiracies as real and refute others as unreal. Indeed, real conspiracies usually do come out; there have been mathematical models showing that the larger a conspiracy is, the sooner it will be exposed. The more people are in the know, the more inevitable it becomes that somebody will accidentally or intentionally reveal what they know. Which is why the fictional conceit of the vast, overarching conspiracy that lasts forever and is never exposed is absurd. Real large-scale conspiracies, like Watergate, Iran-Contra, and so forth, inevitably get found out. It's the conspiracy theories unsupported by evidence that are nonsense.
The problem is that conspiracy theorists don't think in such a rational way. Rather than reasoning from evidence to conclusion, they start with the conclusion they assume is true and interpret the evidence through that assumption. Since they're convinced that a secret conspiracy exists, then they interpret the complete absence of evidence for it as proof that it's being covered up. If multiple sources independently refute the theory, then they all must be in on the conspiracy and it's even vaster and more pervasive than we thought. It's self-fulfilling, circular argument. It's a delusional way of thinking that shouldn't be equated with the rational investigative process that is necessary for exposing the real conspiracies carried out by people in power, any more than UFO cults should be equated with the scientific search for extraterrestrial life, or homeopathic "cures" should be mistaken for actual medicine.
The thing that convinced me most solidly that Lee Harvey Oswald acted alone in assassinating President Kennedy was that at least three different investigations -- the Warren Commission, a CBS team led by Walter Cronkite, and an investigation by PBS's
NOVA -- all independently came to that same conclusion. There's only one right answer and many wrong answers, so if multiple different sources come up with multiple different answers, like the various Kennedy conspiracy theories out there, then most or all of them have to be wrong, but if multiple different sources independently reach the
same answer, that's pretty solid evidence that it's the right one. (That's the whole basis of science, in fact -- any result must be repeatable, arrived at independently by multiple sources to rule out bias and human error.) What's more, Cronkite started out believing Oswald
didn't act alone, set out to refute the Warren Commission's findings, and ended up concluding that they were correct after all. If you set out with a preconception and "prove" that it's right, then your bias could've contaminated your results, so that conclusion can't be trusted. But if you set out trying to prove one thing and end up convincing yourself of its exact opposite, as Cronkite did, that means the evidence outweighed your belief instead of the other way around, and that makes it a very compelling result.