Here's a section of a Gwinn article:
A little math
Of course, in the above analogy, it matters where you point your spotlight. If you aim it at your neighbor's feet, you'd know he was wearing black shoes but you'd miss his red shirt. If you aimed too high, you wouldn't see a thing. The idea is to "center your beam" to illuminate the target to best advantage. With a flashlight you can see where the beam center is, with radar you'll have to calculate it.
Conveniently, the same formula that applies to beam diffusion works with tilt and beam-center calculation: 1- degree of tilt up or down moves the beam center 1000 feet up or down at a distance of 10 miles from the antenna. So, when you nudge the tilt 1 degree, the center of the beam moves 6000 feet (up or down) at 60 miles, 10,000 feet at 100 miles and 18,000 feet at 180 miles. Notice the pattern here? Simply add two zeros to the range and that will tell you how many feet you're moving the beam up or down.
Okay, fine. So what? Now that you know where the beam center is, where are you supposed to put it? I can give you a pretty good idea. In the convective environment that gives birth to thunderstorms, the diagnostic altitudes are 18,000 to 25,000 feet. And what goes up that high, will usually come down. Sometimes with enough energy to bring down an airliner. Thunderstorms go up; non-hazardous rainshowers maintain their low profile. The FL180-FL250 altitudes are where the severe storm symptoms occur and it's where the NWS looks when it's forecasting storm intensity and hazards.
Simply stated, from low altitudes, we'd like to look (tilt) up into that area. From high altitude, we'd look (tilt) down.
Radar Tilt Management