Speaking of spreadsheets, here's my engineering approach.
Wanted to lose weight, 20 lbs from 170, to get my BMI to a more reasonable value, so I bought a digital scale that can show 0.2 lbs difference. I figured I could see how much I gained in a single day while I still remembered what I ate.
Keeping track of the evening weight, in the first 20 days the Excel graph showed a variation of +/- 3 lbs/day in the shape of a sine wave and a weekly upslope on Fridays (because I fell off the wagon).
Also, I lost up to 2 lbs overnight, evidently from respiration, transpiration and elimination.
No food is eaten so this weight change is due only to the factors above. The loss due to the fat burned by metabolism during sleep won't show up for up to 50 hours.
With effort, I stretched out the weekly gain to every two weeks or so but it still went +/- 3 lbs/day and I wasn't losing much so I adapted a modified "Eat to Live" diet by Dr. Fuhrman.
Using a spreadsheet to calculate the trend line slope will bring out the "signal" of -1 or -2 lbs/week from a variation of +/- 3 lbs/day (the signal plus "noise"). This is a very poor "signal to noise ratio", technically speaking.
You can also use a graph and just eyeball the trendline.
Now I'm doing two or three measurements per 24 hr day to further isolate the weight gained from what I know I've eaten, to the noise.
The present downslope is about 5 lbs in 3 days which is a bit fast. One to two lbs/week is more reasonable.
Last edited by Wobbly; 03-12-2014 at 05:03 PM.