I posted this in another thread, but it remains relevant:

"Before I had an incident meter I found the exposure value for the palm of my hand. I metered my hand, then metered a gray card, then compared the two readings. I found that my palm was exactly one stop brighter than the middle gray. I'd go around metering my hand, and then opening up a stop. This technique was almost flawless in shade or soft light, but in bright sunlight the natural oil in my skin caused a reflection that led to underexposure. The technique was also difficult to use in side-lighting. Those factors are what began to get me interested in an incident meter. "

I would bracket in 2/3-stop increments (my camera, a Nikon N90S, had 1/3-stop steps). I got to the point where I could bracket 2/3 over and under, and usually feel fairly confident that I had what I wanted. If the scene was lit by bright sunlight (almost never), I would bracket an extra 2/3-stop over and under. Eventually, I got sick of the hassle, figured that I was wasting a ton of money and that an incident meter would pay for itself, and bought a Sekonic L-508 meter.

Because Provia 100F probably isn't as forgiving as what you're used to, calibrating your system to your preferences could be very, very helpful. I'd suggest running at least one test roll, noting the exposures, later deciding which ones you prefer, then choosing an ISO based on those exposures. With my own system, I rate Provia 100F at 80.

Good luck, and please post any more questions you have about this topic. It's very interesting stuff!