F. Albert Cotton's recent demise brings up a question that traditionally comes up in the fall, during Nobel season. Cotton himself never won the prize, although his name came up constantly in the list of contenders. There's a group of scientists (a select one) in every Nobel-bearing discipline that fills this role. Some of these people eventually get Nobel recognition, of course, and when that happens a good number of onlookers are relieved that ol' So-and-So finally got it, and another host are surprised, because they'd already sort of assumed that ol' So-and-So had received one years before.
But as time goes on, it seems to become clear that some eminent people are just not going to win, and I'd have to have put Cotton in that category. The Nobel committee had years in which to act on his behalf; they never did. The question then is why. Theories abound, some of them conspiratorial (and thus unprovable for another hundred years or so), but most trying to discern what makes some work Nobelish and some not.
One of the strongest arguments is that doing a lot of good work across several areas can hurt your chances. It seems to help the committee settle on candidates when there's a clear accomplishment in a relatively well-defined field to point at. Generalists and cross-functional types are surely at a disadvantage, unless they can adduce a Nobel-worthy accomplishment (or nearly) in one of their areas. That's not easy, given how rare work at that level gets done even when you've devoted all your time and efforts to one thing.
The current example in organic chemistry is George Whitesides at Harvard. He's an excellent chemist, and has had a lot of good ideas and a lot of interesting work come out of his group. But it's all over the place, which is something I really enjoy seeing, but the Nobel folks maybe not as much. Just look at this bio page from Harvard, and watch it attempt to pull all his various research activities under some sort of canopy. It isn't easy.
To drag the late Isaiah Berlin into it again, Whitesides clearly seems to be a fox rather than a hedgehog. Hedgehogs tend to be either spectacularly wrong or spectacularly right, and that last category smooths the path to greater formal recognition. For more on fox/hedgehog distinctions in other disciplines, see Daniel Drezner (international relations), Andrew Gelman (statistics), and Freeman Dyson (physics), and for an application of the concept to drug research, see here. Which sort of creature does Whitesides stock his research group with? Paul Bracher would know.
(Readers are invited in the comments to submit their own candidates for scientists who always seem to be on the Nobel list, but haven't won, and any alternate theories about why this happens).