These techniques seem to work, since I stumble quite frequently upon them in papers accepted at important venues:

• misleading claim — come up with your own definition of a widely accepted term, preferably in a way inconsistent with the intuition and the general understanding of the term. Then claim spectacular results that happen to be true only with respect to your definition.
• marketing pitch — use data irrelevant to the scientific discussion. Bonus: cite a dubious, hard-to-verify source with a non-existent methodology. Ex: “Problem X costs more than Y gazillion $$$ per year (hence our work is of epic importance)”.
• over-sophistication —  use overly sophisticated terms for techniques or ideas that would otherwise be simple to understand (especially if it makes them look better). Ex: “heuristics” [1] sounds a lot better than “hack”.
• no limitations — there are no limitations to your work, only things that are either out-of-scope or left for future work.
• outsourcing problems — outsource large portions of the hard problems in your area by “leveraging” (see over-sophistication) existing work from a different domain with a complex literature, just to make sure that nobody can grasp all the dependencies of your paper.
• biased evaluation — perform experimental evaluation designed to make others look bad. Even better if the evaluation is not reproducible, because you fail to disclose your source code, detailed algorithm or input set.

[1] heuristics (noun — hyu’-ris-tik): algorithm that does not work.