Personally, for longer papers (20 pages+), I think a thesis statement is not the correct way to go about things. I think something like a thesis needs to exist; that is, an explanation to your reader as to the topic of your paper, early on in your paper. However, I think the "it must be a single sentence" constraint is unnecessarily limiting. Particularly in technical fields, an explanation of one's position may require several sentences, or even a paragraph, in order to be clear.
More to the point, I've never read a professionally published work where a "thesis statement" in the traditional sense was obvious. When you're writing papers about political issues or issues which are particularly contentious, then it makes sense to have a "line in the sand" which describes which side of the debate you are on. But most academic work I've done/seen is not work where there is an obvious position one can take.
Finally, I think that even in cases where a line in the sand must be drawn, I like a structure where the topic is explained early on in the paper, and where you manipulate the reader to come to the the conclusion you want, rather than drawing that line manually. An example of a paper that does this is A Climate of Belief -- and it's done quite well. (There is a thesis statement of sorts in the tagline here, but it's not worked into the body of the paper)
Therefore -- I don't think thesis statements make sense as an end-all be-all solution (as the are purported to be by English teachers and professors almost universally). Am I wrong here?