Re: OK I admit it
"thought about publishing the data somewhere? "
Oh, it's almost all available if you have good access to the the biomedical literature. The only thing I really struggled with is the value for height gain as there seems to be a wide range in the literature. My final value of 100 cals/100m gain+loss and matches my actual loss rather well.
Another interesting on-line source , esp to an IT audience, is a document by John Walker one of the Autocad founders although he lost weight mostly by calorie reduction.
The Hacker's Diet
How to lose weight and hair through stress and poor nutrition
By John Walker
Unfortunately I'd worked out my routine and lost a lot of weight before I found this.
"coming in from a 2.5km run and eating 3 slices of toast and honey - vastly more calories than I have theoretically burned."
You can't treat it like that. You need to know at the minimum how many calories/day will keep your weight steady. Is the toast etc in addition to that basal calorie level. There really is no magic in this for most people long term unless you are one of the 'lucky' ones with lots of brown fat (assuming it really exists/works in humans) or some metabolic issue ( malabsorption of fats for example)