Although I don't count calories to maintain weight, if I ever want to lose weight I have to. I use MyNetDiary to track everything. Just simply thinking about stuff for entry usually has enough of an affect on my eating patterns.

As far as how much is burned through biking, the most accurate way to know is through a gas exchange respirator. Obviously that's overkill and not practical. A power meter is probably the next best, especially if you have done a gas exchange test in a lab to determine your efficiency. However the range of human efficiency is pretty narrow so this is not critical. Estimating power has it's own problems, but is probably still going to be better than most x speed = y calories calculators. HRM calorie counters can be ok but are heavily based on statistics which you may or may not match. The better ones take into account your resting HR, max HR and an approximation of your VO2. And of course your weight, age, etc. The absolute worst way is to use those generic exercise amounts that come with the food tracking sites.

Note, for the following, assume 1kJ of work (measured by a power meter) is roughly equal to 1kCal of food energy burned.

As for any kind of x calories per hour or y calories per mile, those will depend very much on how much you weigh and how hard you are working. For example, I did the same loop yesterday and today, but yesterday was a recovery pace where I did 383kJ of work in 50 minutes, and today was tempo where I did 513kJ of work in 40 minutes. Both were 14.1 miles, so yesterday was 27.4kJ/mile and 460kJ/hr but today was 36.4 kJ/mile and 770kJ/hr. That's a big difference. This was a relatively flat ride, the amount of climbing will decrease the amount per mile (for the same effort). And if I am really working hard, I can do over 1000kJ in an hour. And I'm a small guy, bigger people will burn even more because it takes more work to go the same distance. The bottom line is any kind of generic calories per mile or calories per hour will assume too much.