Is there a limit on how low 3x3x3 times will get?
For single times, the limit doesn't exist. How low times get is influenced in large part by how lucky the solve is.
Average of 5 is better, but IMO suffers from the same problem. It's not very likely that you'll get three-four OLL skips in a row (or some such), but it only takes one world-class cuber to do so and suddenly you've got a very low ave5 WR on your hands. Honestly this isn't a problem yet, but if cubing sticks around for another 50 years we're going to see lucky averages come up much more often.
On the other hand, there's definitely a limit on how low times will get. We're just not there yet. We'll know when we are, as we'll start seeing WR times beaten by smaller and smaller increments, and WRs will start to get less common. They'll also start to get more lucky, which ties in with my paragraph above.
I think it's important to look at older, more established contests for comparison, and I think timed sporting events do a nice job. (100 meter dash, for example). Generally the world records in these events fall very quickly, then on a regular basis, then fairly rarely. Meanwhile, the amount that the records decrease by goes down. When I was in school at some point they showed me a graph of world records in some such event versus time with a best-fit curve showing how each record seemed to be getting closer and closer to some limit. I'd imagine that cubers in 50 years will be able to do so as well.
EDIT: Tried it, but couldn't get it to quite work right. (Using open office instead of excel and I'm not really familiar with it.) Looks like we don't have enough data in any case; 14 is a low number, especially since the first few are much higher than the rest and two happened on the same day. However, I'd be interested to find out if someone who's better at statistics can come up with a reasonable asymptote...
