Hanks theory is very likely correct if I understand it right... He seems to be saying that when you draw alot higher amps then the packs were rated/matched at, then your calculations aren't going to be right if you use the typical methood... well in fact anything short of testing your batteries capacity at what is very close to your race average is going to yeild less then accuate data...
I do belive it's common for stock oval racers to average at least 30 amps during a race... I'm not sure anyone racing these days ever goes as low as 20 amps, exept possibly for 8 minute 12th scale racers.
If your packs capacity/run time is calculated using 30 amp discharge rates, you should be able to take the tested capacity, subtract the run time remaining (when also done at 30 amps), then devide the number of seconds mAseconds by the number of seconds you raced... That should get you pretty close... using 35 amp discharg numbers will likely get you even closer when your averaging near 35 amps... Start averaging much over 35 amps, and you get back to the problem hank seemed to be pointing out...
There is another solution to all this... get yourself a eagle tree data logger, and collect the amp draw data as you run...