What is the point of having a fast and slow probe cycle?

I've never really understood the point of the initial fast probe. Is it just to get a rough estimate of position, or does it do something more?

Your probe axis may have to run a long distance to touch the probe, so running fast it finishes faster.
The second slow cycle is to get a better accuracy measurement.

That's pretty much it: to get the rough position. The fine (slow) probe is where it actually does the measuring.