Sleep: A Field Report From Year Four

I have been a father for four years and three months. In that time I have had, by my estimate, eleven genuinely good nights of sleep. I remember most of them individually.

Ellie cracked at eighteen months. Not perfectly — there were regressions, the two-year molars, a chest infection that set everything back six weeks — but broadly, by eighteen months, she was sleeping through with the consistency of someone who had decided that sleep was, on balance, a good idea.

Sam is two. Sam has not made this decision. Sam is operating under the belief, seemingly sincerely held, that the night is a resource to be used rather than a period to be unconscious through. He wakes. He wakes with opinions. Last Thursday he woke at 2:17am, stood up in his cot, and announced "biscuit" with the clarity and conviction of a man arriving at a conclusion he had been working toward for some time. There were no biscuits at 2:17am. This was not well received.

This is the field report from year four. It is not a success story. It is an ongoing document.

What actually worked with Ellie

I am going to tell you what worked for us, with the very specific caveat that there is no universal solution, children are not machines, and anyone who tells you there is a guaranteed method is either lying or describing a coincidence.

With Ellie: consistency over everything else. The same routine, the same order, the same words at the same times, night after night, until the routine itself became a cue that sleep was coming. Bath, milk, two books, song, lights out. For three months this produced screaming. Then gradually it produced less screaming. Then it produced a child who started asking for the routine before we initiated it, which is when you know you've won.

We used a white noise machine. This is one of those interventions that feels too simple to actually work and works completely. It masks the ambient sounds of a house that is still awake — the television downstairs, the boiler clicking, the sound of Claire and me attempting to have a conversation at a volume that won't carry — and it provides a consistent audio environment that a child's brain starts to associate with sleep. We use the Hatch Rest. It doubles as a nightlight and a "okay to wake" clock when they're older. The "okay to wake" feature alone has been worth the price. There is a link in the Dad's Toolkit for when it becomes relevant.

What is currently happening with Sam

Sam is at a stage that the parenting literature euphemistically calls "night waking." What this means in practice is that he wakes, he wants someone, and the want is expressed with a volume and persistence that is not compatible with leaving it to resolve itself.

We have tried the consistent-routine approach. The routine is solid. The routine is excellent. Sam enjoys the routine and then wakes up at 2:17am anyway.

We are currently in the "reduce the number of wake-ups" phase rather than the "eliminate them" phase, which is the realistic goal with a two-year-old who is cutting teeth and going through a developmental leap simultaneously. The Hatch is earning its keep. Claire and I take alternate nights on the Sam duty. The nights you are not "on" are not good nights — you still register the sound, you still half-wake — but they are meaningfully better than the alternative.

What the older body does with all of this

Covered in detail elsewhere, but the short version: recovery takes longer. The same disruption that a 31-year-old might absorb in one good night costs me two or three. I am not complaining about Sam. I am noting the physiology.

What I have found useful at this stage of the process:

Magnesium glycinate before bed. This is not a miracle supplement. It is a very boring supplement that helps slightly with sleep quality and has no side effects worth mentioning. A former colleague who is a GP recommended it. I was sceptical. It made a measurable difference to the Garmin data. I remain sceptical in principle while accepting the evidence in practice.

Not looking at the time when Sam wakes. This sounds trivial and is not trivial. Knowing it is 2:17am is worse than not knowing. The specific number does something to the brain that prolongs the return to sleep and adds a calculation about how many hours remain. I turned the clock face away. It helped.

The commitment to the long view. Ellie cracked at eighteen months. Sam will crack too. All children eventually sleep through. This is empirically true. When I am at 2:17am with a child demanding a biscuit, the long view is not my immediate comfort, but it is available as a background truth.

The thing about sleep that nobody puts in the parenting books

The parenting books treat sleep as a problem to be solved. Get the routine right, apply the method, achieve the outcome. This framing is understandable and occasionally useful and also slightly missing the point.

The night wakings are not purely a logistics problem. They are also the times — the 2:17am times, specifically — when you are alone with your child in the dark and there is no performance and no audience and nothing required of you except presence. Sam, who demanded a biscuit and then leaned his full toddler weight against my chest with the confidence of someone who has decided I am a reliable surface, settled back to sleep in about eight minutes.

I sat with him for another five.

This is not efficient. It is also not something I regret.

The field report from year four: hard, ongoing, occasionally luminous in the specific way that 2am can be luminous when no one else is awake.