Noise, and the line through it

Everything so far has said the same thing from different sides. The number on one morning is noise, and the only thing the scale can tell you is the line. This chapter is about what a line looks like, so that when you see one you'll know it, and when you don't you'll know that too.

Take your sheet from chapter 1, if you kept it going, or the last few weeks of readings if you've got them anywhere. Write the numbers in a column with the dates. Now, instead of reading them one at a time, take each week's readings and average them. If you weighed seven times, add the seven and divide by seven. Write the weekly averages in a second column. Look at that column and only that column.

What you'll usually see is that the daily numbers jump about, up half a kilo, down a kilo, up again, and that the weekly averages hardly move. That flat column is the line, and it's telling you that over the period, nothing much happened to your weight, whatever the mornings felt like. Orsama's study from chapter 4 is a picture of exactly this. The Monday numbers looked like gain and the Thursday numbers looked like loss and the week as a whole was going wherever it was going, slowly.

Sometimes the averages do move. Three or four weekly averages in a row going the same way, each a little further than the last, is a line, and it's real. That's the signal, and it's the thing a smoke alarm should be for. It takes about four weeks to see, in my experience of looking at these columns, because a single week can be pushed around by the time of the month or a holiday or an illness, and it takes a few weeks for that to wash out. Four weeks is the figure I'll use through the rest of the book. It's a judgement, not a finding.

The one analogy for this chapter. A tape measure is a good instrument, and if you measured your waist over a winter coat you'd get a true reading of the wrong thing. The scale over a curry is the tape over the coat.

To get a line worth reading, the readings need to be taken the same way each time, which is why you already weigh before the kettle. Same time, same state of dress, after the loo, before anything else. Same scale on the same bit of floor. That doesn't make any single reading mean more. It makes the readings comparable, so that the average is an average of like with like.

Once you can see the line, the number on any given morning loses its grip, because you've watched a whole column of numbers that looked like judgements turn out to be nothing at all. That's the understanding this book is trying to give you. The next chapter is about what to do with it.