Well, I honestly thought we'd have more discussion here. I'll go ahead.
I think that the idea of repeatability is important in terms of replicating one's results, to verify their accuracy. But that doesn't necessarily mean that if you can't repeat an experiment that it was never science to begin with. It just undermines the veracity of the results. Science doesn't equal "truth." It's a method. You might be able to get true results just once. It's conceivable that some experiments can never be repeated.
Science is contrasted with superstition, religion, and mythology, but also common knowledge, tradition, authority, rules of thumb. It's not just naturalistic (rather than supernatural) but also rigorous, tentative, questioning, and skeptical. It originally stressed empirical evidence as a source of knowledge--contrasted with authority or ancient texts--and this was successful, but it was also misleading. Knowledge doesn't come from empirical evidence. All empirical evidence is "theory laden," as Popper says. We interpret all evidence according to theories that we already believe, as well as theories we want to test. So there isn't really "objective evidence." Not only do we see what we want to see, but we often only find what we are looking for.
Knowledge, as Deutsche says, comes from conjecture, which is then tested against the empirical evidence. But most of the things which we have knowledge of are things we'll NEVER observe. We have an excellent idea of what is happening inside stars, though we'll never see their interiors. Most of our empirical evidence for things we hold true is indirect evidence.
A good explanation is one that can't be easily varied without impacting its explanatory power. The more specific and invariable it is, while still explaining the phenomenon, the better. The more you reveal that "it can't be another way," the closer you are getting to reality.
But even this is not enough. It must also conform to previous knowledge. This isn't a hard and fast rule, because as Kuhn showed, science moves through a serious of paradigm revolutions, in which large bodies of previous knowledge are overturned or reinterpreted. But the "normal" times of science are characterized by working within a paradigm, working out all the details, chasing down the implication of what is already assumed to be true. So something that fundamentally contradicts the rest of science, is suspect. This doesn't mean that it's not true, or not science, but it's a valid reason for doubt.
Another point bought to us by Popper is the falsification principle. If a scientific theory is not falsifiable in principle, then it cannot be science. This is true for many supernatural ideas, though not as many as you'd think. For instance, if we can show that the universe can spontaneously burst into existence due to principles of quantum cosmology, we can in fact falsify the idea that it requires supernatural creation in order to get the ball rolling. So, creationism is falsifiable! That doesn't mean that creationism is a scientific theory. It just means that falsification alone isn't sufficient to tell the difference.
Pseudo-science often depends upon the ignorance of the audience, "lying with statistics," confirmation bias, misunderstandings of probability theory, and many other psychological reasons for people not to think critically/rigorously. Pseudo-science is often much harder to spot than mere mythology, because it dresses itself up in the language of science. But if it is radically disconnected from accepted science, is not falsifiable, contains logical flaws, contains bad math, can be varied substantially without affecting the theory, then we're dealing with pseudo-science.
That's off the top of my head. Now, after Googling to refresh my memory:
Examples of pseudo-science:
Chiropractic (I've discussed this before)
Ley Lines (attributing intentional significance to the placement of ancient monuments that was actually random chance)
Perpetual motion machines
Flood geology (ascribing most of earth's geology to Noah's flood)
UfOlogy
Crop circles
Electromagnetic Hypersensitivity
Magnet therapy
Myers-Briggs personality indicator test
And many more. Very interesting list.
Here's a link to Stanford Encyclopedia of Philosophy on the issue of science vs pseudo-science. This link is fascinating in its complexity, showing the breadth of debate and thought on this issue, including the possibility that the line between science and pseudo-science could move as science itself progresses. Others resist this, insisting that the criteria must be timeless.
An interesting point of emphasis is
science denial. Many pseudo-science theories have as their chief objective the denial of standard, accepted science, such as climate change skepticism and intelligent design. (I disagree that being skeptical of the future impact of global warming to be pseudo-scientific; I think that such projections rely upon Malthusian naivete of the ability of humans to adapt to change--an inherently unpredictable phenomenon in its specifics, but reliable enough generally to be taken into account as a mitigating factor to global warming alarmism.)
It is important to note that many of the features of pseudo-science are part of science. That's why it's so difficult to delineate. Was Einstein's denial of the implications of quantum mechanics scientific? The way he phrased it most certainly was ("god doesn't play dice . . . "). But the way he investigated this skepticism led to the discovery of things like quantum entanglement. So why can't denial be a valid part of moving science forward?
Anyway, if you guys have any pet theories (like chiropractic or global warming skepticism) that you want to defend as scientific, go for it! Likewise, if you want to attack something as unscientific (e.g. simulation theory), go for it! This is the place to get that out of our systems. But let's try to keep it in the framework of the philosophy of science, rather than over-reliance of metaphors/similes that emphasize absurdity alone.