Why it's always the work that matters
Here's an odd thing about your working day. A colleague asks you to look over their slides, and you do it within the hour and do it well. An expenses form gets filled in. A difficult customer gets a careful reply. Meanwhile the one piece of work that could affect your year sits untouched. If the problem were a general inability to get on with things, the slides and the form would be waiting too.
The difference is what the result can say about you. Nobody's opinion of you hangs on an expenses form. The proposal will be read by people who'll form a view of how good you are. Worse, it will be read by you. A piece of work that matters is a measurement, and the swerve is what you do when you'd rather not be measured today.
This explains some things that look strange otherwise. The feeling is strongest at the very start, when the page is blank and the work could still, in theory, be brilliant. People who are good at their jobs are especially prone to it, because they have a reputation, with others and with themselves, and every new piece of work puts it at risk. And the dread is often worst for the work you most want to do well. Steel's review fits this. Two of the strongest predictors of putting a task off were how unpleasant the task felt and how much the person doubted they could do it.
The next part is uncomfortable, and I'd ask you to read it without turning it into a stick. Starting late doesn't only postpone the measurement. It spoils the measurement, in a way that protects you. If the proposal was written in one night and gets turned down, that says nothing about your ability. You know what you could have done with proper time. If it was written in one night and gets accepted, you look better still. Either way the real question never gets answered, which is how good you are when you give something everything. Psychologists call this self-handicapping, a term from Steven Berglas and Edward Jones in the 1970s. You arrange an obstacle in advance so that the result can't be read as a verdict.
It's a bit like a runner who always turns up injured. His times never count against him. They never count for him either.
Joseph Ferrari and Dianne Tice tested the idea on procrastinators in a laboratory study published in 2000. As the study is usually described, people were given time to practise before a maths test, with games in the room. When the test was presented as a measure of their ability, the chronic procrastinators practised less than other people and played more. When the identical task was presented as a bit of fun, the difference went away. I haven't been able to read the paper itself, so treat the detail with care. The pattern, though, is the one you've just recognised in your own day. It wasn't the maths they were avoiding. It was the measurement.
I don't think anyone does this on purpose. Nobody thinks "I'll start late so that my ability stays unmeasured". The swerve takes a second and feels like nothing more than a preference for tomorrow. The protection is a side effect that happens to be welcome, and welcome side effects get repeated.
You may be thinking that this isn't you, that you aren't afraid of failing and the task is simply tedious. It may be. Some work that matters is also boring, and the feeling you named in chapter 1 might truly have been boredom. But check it against the expenses form. You do plenty of boring things promptly. If the boring task that waits is always the one with your name on the front and a judgement at the end, boredom isn't the whole story. The same goes for confusion about where to begin. When a result doesn't matter, you're happy to begin in the wrong place and sort it out as you go.
What the late start protects is an idea, the idea of how good you'd be with proper time. It's a pleasant idea to own. The price of keeping it is that you never find out, and you improve more slowly than you could, because improving needs honest measurements to work from. Ten years on, the idea is the same size as it was, and the work hasn't had the second passes that would have made it better.
If the thing being avoided is judgement, then a method that works has to take judgement out of the first session. Trying harder doesn't do that, and neither does a schedule. A start that's supposed to be bad does, and that's where we're heading.