Science isn't a computer program based on some simple algorithm. I understand that it can be difficult to spot crackpot to a layman.
1. String theory doesn't assume 20 crazy things and its axioms aren't its only predictions
2. String theory does have a basic equation from which you draw many results, leading to falsifiable predictions that is not covered by standard model. Some of them may not practical with today's experimental tools, but that's not all of them. I take you haven't seen any string theory papers. I'm saying we're suddenly going to see superparticles this year at LHC, but it is entirely plausible that if they exist, LHC should eventually detect them.
3. You're probably talking about this http://advances.sciencemag.org/content/2/2/e1501466 recently posted on HN. In case you missed, that is an experimental paper.
