Silverbirch, I weigh daily Monday-Sunday and take the 7 day average. Like everyone else, I get the unexplained blips and the temporary "whooshes". As an engineer who works constantly with statistics, I am immune to what happens on a day-to-day basis. A single daily data point is meaningless to me - I'm numb to them. What does catch my attention is an average based on a properly obtained sample (same time every morning with all the same conditions) over a long period of time.

I don't make graphs of it (I used to), but a control chart essentially follows an average over time, with a +/- "buffer" based on standard deviations. With a control chart, there are strict rules for interpretation based on the sampling. This is the sort of control that makes me the most comfortable and it helps separate a true trend from "noise". My weight averages were, literally, out of control on the upside, which meant it was time to adjust the process. Hence, the "Trust the Process" in my profile.

If anybody really wants gory details about control charts, they are here:
http://www.wikihow.com/Create-a-Control-Chart . Instead of using a - d in the example, I use Mon/Tues .../Sun. 20 "subgroups" is 20 weeks of data, which I actually do have, back from 2009 when I lost the bulk of my bulk. A control chart is well suited to Maintenance - it's the theory behind the practice of the 5-bar maintainer's ticker that Bright Angel uses. Are you sorry you asked yet?
