That's comparing multiple people with positive favorability ratings... your statement said that someone with a -25% favorability rating is more likely to win than someone with a +25% favorability rating. That's so laughably absurd I don't even know how to argue against it.
So?
You presumably think Romney is going to win, yet Romney's favourability is lower than Cain's. Ergo, using your logic, Cain will definitely win the nomination.
Remember, not too long ago Cain had negative favourability whereas Romney had positive. Does this mean Cain suddenly had no chance?
Favourability changes at the drop of a hat from poll to poll and you have yet to demonstrate that it effects the end results in any significant way. Way back when, Paul had around +30 in New Hampshire, just a tiny bit behind Romney. He had similarly high ratings in Iowa. Every other candidate had similar sudden drops and increases because unlike direct polling numbers, favourability is rather mushy. Joe Republican might consider Romney favourable when listening to him talk, and then consider him unfavourable after hearing that he was strongly in favour of abortion not so long ago. Similarly, he might like Perry's posture and that he is Texan and consider him favourable, or he might hear him actually talk in a debate and consider him unfavourable.
It doesn't even have to come down to "bad sides". One voter might be hearing about one particular issue and thus like the candidate and then dislike the candidate when he changes topics. Paul would be a fine example of this; When he talks primarily about government spending, states' rights, etc the Republicans suddenly like what they hear and decide to increase his mark him as favourable (without voting for him, of course). Then he talks about foreign policy primarily and they find it strange to be coming out of a Republican's mouth so they mark him as unfavourable. That isn't even mentioning, for example, media coverage, spin, etc (The guy who is being covered by news channels constantly is going to see the bigger fluctuations than the guy who isn't). A Ron Paul given a single week of constant media coverage for his trillion dollar cuts exclusively would do to his favourability what a week of constant media coverage of Perry or 9-9-9 did for them. A Cain given a single week of constant media coverage on foreign policy or a Romney in a debate with someone pointing out his abortion/gun rights views would both see similar drops.
Favourability has no relevance at this point.