I'll admit I'm a stats junky. I keep track of feet/mile on most rides. For the really hilly rides, I keep track of the average climb gradient and the max gradient (I'd do it for more, but I really, really detest garmin's website). I also like to track the number of hills that exceed 10% as this seems to really correlate with the pain in my legs.

Interestingly, I've found that the simple feet/mile "adjusted for statistical error" does give a pretty good indication. Ties come through with the max gradient and # of climbs over 10%. Not to start a "my hills are worse than your hills" war, but I find that a long, sustained climb of 4-6% is always easier than rollers with 0.5 to 1.5 mile climbs that hover near 8-12%. It is just me; others believe otherwise.

The guy that runs the Dairyland Dare has a formula where he takes the feet/mile and adjusts for distance. ( www.dairylanddare.com ) That doesn't cover SouthernBelle's argument....if it is a flat route with one killer climb.