Not going to get into all the arguments here. Yes, it is more complicated in detail than the simple model Walker lays out. But in practice, *if* you count calories as prescribed, *then* the model is good enough.
I'd like to provide an update here. I read about the Hacker's Diet first on Slashdot, in fall 1999. I followed it, and during 2000 I lost 50 pounds. I've kept it off for 13 years now. A few years later I started running. I've now run 96 marathons and ultramarathons, heading towards my 10th consecutive Boston Marathon, I've broken 3 hours four times, and I've run three 100 milers, including Western States. Couldn't be happier with that part of my life.
The running has been a bigger life change than losing weight. But I couldn't have done it, no way, without losing the weight first. And I have the Hacker's Diet to thank for that.
And yes, running 60-70 miles / week, I *still* have to count calories.