My political predictions are somewhat better than my civil war predictions, so with that faint praise, here's one: if Obama doesn't approve Keystone before November, then he won't approve it at all. It's good politics but bad policy to approve Keystone. The political reward diminishes after the lower-turnout, more conservative, off-presidential election year, while the policy and international reasons for not approving Keystone get stronger with more delay. And not to get too wildly hopeful, stringing along the idiotic Canadian government before cutting them off might make it harder for PM Harper to figure out alternative ways to harm the planet with tar sands.
I only put a little emphasis on the recent addition of climate hawks and Keystone opponents to the Obama admin - he's the one they work for. It might be a hint of how the wind's blowing, though.
OTOH, there's still 10 months to make a decision. And just to caveat it a bit, I could imagine Obama locking himself so close to a promise before the election without getting to the final signoff that he does approve it soon afterwards. So far though this reinforces the obvious point I made almost two years ago that delay is a victory (albeit not THE victory) when you're trying stop somebody else from doing something. Just like it is in Iran.
The other prediction came from reading in Stoat of the welcome retirement for Richard Lindzen, leaving the ranks of non-retired, denialist climatologists even thinner than before. John Mashey noted in the comments that the 2009 "mass" petition to the American Physical Society to deny climate science from 0.45% of its membership was skewed much older than the general membership. My second prediction is this petition will not be repeated because it will be even less successful in the future than it was in 2009, as demographic destiny wins out. This might explain why Singer and friends kept switching from one meritless type of mass document to another - recycling an old one some years later will just highlight increasing consensus.