Finding out how long things take

The first practical step is to replace the estimate with a record. The estimate comes from a picture and leans towards the best case. The record comes from the clock and doesn't lean anywhere. Chapters 3 and 4 gave the research. This chapter is the method, and the method is mine.

If you did the exercise in chapter 1, you have a week of predicted and actual arrival times. Look at them. I'd expect two things. The actual times will be later than the predicted ones nearly every time, and rarely earlier. And the gap will be roughly the same size across quite different trips, because the gap isn't about the trip. It's about the method you used to estimate it.

Now time things properly for two weeks. Don't try to be quicker. The point is to find out what's true, and if you speed up to look good, you'll get a record that's as optimistic as the estimate. Time the trips you make often, and time them from door to seat. Not pavement to pavement. From the moment you decide to leave until you're sitting where you're meant to be. Write both times down. Do the same for the things you do before leaving, if you can bear to. How long is a shower, really, including drying your hair? How long is getting a child into a coat and shoes? How long is finding things?

At the end of two weeks you'll have, for each regular trip, a handful of real durations. Take the longer ones, not the average. If the walk has taken fourteen, sixteen, twenty-one, seventeen and twenty-three minutes, the number you want is somewhere around twenty-one. That's the trip on an ordinary bad day, and ordinary bad days are, by your own account in chapter 13, ordinary.

This will feel pessimistic. Newby-Clark's study found that people rate the version where things go wrong as less plausible than the version where they go right, and you'll feel that pull. Twenty-one minutes for a fifteen-minute walk will seem absurd. The record says otherwise, and the record was there. This is the outside view from chapter 13, applied to your own life, and it's the only method I know that produces a number you can actually leave by.

For trips you haven't timed, unpack them. List what the trip involves from where you're sitting, including the finding and the locking and the stairs, and put a minute or two against each part. Then add the cushion, which for most trips I'd set at ten minutes, and for trips involving traffic, public transport or children at something more like twenty. You'll find your own figures. What matters is that they come from counting, not from picturing.

Once you have the number, use it without argument. The moment you catch yourself thinking "It won't take that long today," you're back in the picture, and the picture is what got you here.