Sure... but he must win of every one of
Florida
Missouri
Nevada
North Carolina
Ohio
Virginia
One chance in 64, which is a gamble that he can't find adequate. No single appeal or issue wins all of them.
In theory he could win Michigan or Pennsylvania as a substitute for any one of those states...but I wouldn't bet against a strong union GOTV drive.
But that isn't one chance in 64. You have to adjust for details -- Romney is more likely than Obama to win MO, the opposite is true in NV. And in your last sentence, you forget once again that Wisconsin is clearly showing the anti-union drive is more powerful than the union drive nowadays.
We shall see on Wisconsin soon enough. There is no good model for predicting how a freakish off-season special election goes. The Hard Right is flooding the state with ads saying that if 'anyone but Walker' will quickly have people sorry to have so voted. When logical reasoning fail, use threats, as some fundamentalist preachers use the "Believe it or Burn (in Hell)" argument.
The "(1/2)^n" model is easy to apply to coin tosses... but of course the chance of President Obama winning any one of those states is clearly not one chance in two. For President Obama it's
1- (1-a1)(1-a2)(1-a3).... (1-a6) for random chances with independent events with each "a" (I can't put subscripts in the model) here representing the probability of a Romney win in any particular state. But here's one general fact:
If the chance of one of those states going for President Obama goes to near 100% then the chance of Obama winning no longer depends on what happens elsewhere. So if the chance of Barack Obama winning Virginia goes to .99 and his chance of winning Missouri completely vanishes, then a 50% chance in all other states pushes the chance of an Obama win to 99.9375%. We just might see something like that happen. Time narrows tangible possibilities in something full of random possibilities. Someone batting .213 on April 17 might still end up batting .300 at the end of the baseball season, but someone who has played regularly and is batting .271 on September 28 has practically no chance of hitting .300.
As it is President Obama now has a tangible chance of winning Texas, and Mitt Romney has a tangible chance of winning New Jersey. By late October the chance of either happening will most likely be nil.
