Nice. That's pretty awesome. Makes a lot of sense if you think about it.
However, I never got the impression in my study of quantum electrodynamics that photons actually hold more probability of going back in time. Processes, however, would be invariant under a time shift (specifically, a CPT--charge, parity, and time transformation--would be invariant). http://en.wikipedia.org/wiki/CPT_symmetry http://en.wikipedia.org/wiki/T-symmetry
In the original article you linked, there was a paragraph which said "The same holds true in a less dramatic fashion for subatomic particles like quarks and photons, which hold a much higher likelihood of ever traveling into the past, based on the strange, spooky behavior they’ve demonstrated for the people who’ve studied and been alarmed by them thus far." That really seemed... incorrect to me, so I clicked the link, and it just says a bunch of stuff about the Many-worlds interpretation, which is just an interpretation of quantum mechanics. It is not the orthodox interpretation (at least, according to my wonderful quantum mechanics textbook). This does not mean that particles are actually going back in time, and I think was extremely poorly worded.
When you look through the topic on the Many-worlds interpretation, ugh... I'm in quantum mechanics right now. That article did nothing to imply that particles are going back in time. This is why you don't get your information on How Stuff Works, or pretty much any popular science article. *huff huff* You don't really start to understand quantum mechanics without a lot of math. That's just how it is.
See, you have to realize that the equations sort of allow you to do a time shift. If you give me a wave function, I can tell you its amplitude (think probability) at any time t, positive or negative. Just because I can write it down does not mean that the particle actually moved back in time. There's nothing mysterious about time travel in regular quantum mechanics. But that's not even mysterious in non-quantum systems. For example, I can write down that the distance a car moves across, I don't know, an interstate, would be d = vt. I can plug in any time that I want, positive or negative. Does that mean the car moved back in time? No. Obviously not.
Now, happyman, you know more physics than me, but I got a good taste of quantum electrodynamics. I am fairly sure "subatomic particles like quarks and photons, which hold a much higher likelihood of ever traveling into the past" is a false statement. Maybe there's some weird quantum field theory which actually states that, but I always interpreted it as that the math doesn't particularly care which direction time is moving in the process.