The following analogy is presented as a partial solution to the problem of finding a classical version of the Uncertainty Principle.
I hit upon the basic idea while leaving a movie theater as other people were trying to get in. I gave wide berth to an old woman with a cane, but all but tried to ram a group of highschoolers that were in my way. Then I stopped just outside the crowded area and realized I'd just mentally computed probability distributions...at a Classical level! I'd made sure the chances of running into the old woman were very low, but was willing to accept fairly high odds of hitting a 9th grader. And thus I had an analogy.
Few students have had direct experience with quantum effects, but they've all had the experience of trying to navigate a hallway or theater entrance full of people. Person A needs to guess where Person B will be in a few seconds in order to avoid a collision. Person A will almost automatically guess how likely any given location is for occupation and set limits on how close he's willing to cut it. Probably at that point, pretty definitely within that general area, 100% positively within that larger area. The person has just set up a probability map without even thinking about it. Helping students actually start to think about what they're doing in this sort of situation can go a long way towards helping them develop an understanding of quantum probabilities.
To further illustrate this point, two special cases will be considered, and analogies drawn to orbitals in a Hydrogen atom. After this has been done, mass and time will be discussed as a means to extend the everyday experiences down to the quantum level.
However, one second from now she may no longer be standing there. She could have started moving in just about any direction, or shifted around slightly to look at something. You can assign a high probability of her being in the same place one second from now, but not 100%. Given a relaxed walking pace of about 1 meter per second, you can draw a circle 1 meter in radius around her current position and say with even higher certainty that she'll be inside that circle one second from now. Perhaps she'll start moving at a more urgent speed, around 2 meters per second, giving the circle of 2 meters radius a slightly higher probability of containing her. And so forth, until you have to invoke wildly improbable events to get her farther out in one second...but not totally impossible events. A graph of her distance from the starting point at one second from now might look something like the following graph.
Now, remove the knowledge of where she is now and keep only the graph of probabilities. You now have a map of possible locations for a classical object at a fixed time, much like the probability plot of a ground state hydrogen electron (although not exactly like it). No particular direction is preferred, and you can't be totally sure the person or electron is at the mean radius or location.
Consider the woman now moving down the hall at a fairly steady pace. If she keeps going the same speed and direction, you can say where she should be in one second. However, it's possible that she'll speed up, slow down, stop entirely, turn or even do an about-face. A contour graph of her likely positions will resemble a bowling pin contour map, with two oblong regions joined together. The more likely and larger region will be centered on where she would be if she stayed on course, the second oval would center on her current position, assuming a stop and then random motion from there. A graph of her velocity one second from now would have a similar double hump, one around her current velocity and one around zero.
Obviously, the metaphor starts to seriously break down once you reach the f-orbital and higher angular momenta, but hopefully s-orbital and p-orbital analogies should be enough to help students grasp the idea.
Now look at the other direction. Replace the woman with an elementary school child half her mass. The child is less likely to stay in one place, and more likely to run about at random. Knowledge of either position or velocity for the child now will tell you very little about those quantities one second from now. Graphs will flatten out, it will be harder to predict anything about future activity. Extend this further, all the way down to a butterfly, which is very light and very unpredictable when in motion, moving sometimes almost at random. The lighter an object gets, the harder it is to say anything with certainty.
Now, while the analogy gets a bit weak as you move outside the range of human masses (a snail's small, but pretty certain, for example), the following relationships should be fairly acceptable to students: