I've added the regression line. The R-squared statistic for this data set is the square of the correlation coefficient, so the correlation between weight and elapsed-time is around 0.7. This shows a real trend.
The bouncing around of the data from the trend line reflects days of eating and days of fasting. Food taken on board during an eating day can weigh of the order of 1.5 - 2 kilograms. This gets mostly metabolised into water, carbon dioxide and solid waste over the next day or two. Hence the jagged shape to the data based on fasting two days a week. The data tends to cluster within +/- 1 kg of the trend line.
I have a target: be at a mean weight of 68 kg = 10 st 9.6 lb and wobble by around +/- 2 pounds during a typical week. The trend line says I will get there on September 6th 2013.
|My weight (kgs): Jan 1st - June 15th 2013|