The peak is not where anyone tells you it is
2026-08-25
The worst stretch of nights was not the newborn period. Across 670 hand-logged nights of one child, wakings peaked at seven to eleven months — a mean of 3.5 a night, and as many as eight — comfortably above the 2.2 of the first two months. Then they fell back. I had that data for over a year before I thought to plot wakings against age; the hump was sitting in my own logs the whole time.
The shape isn't mysterious once you line it up against everything else happening at that age. Separation anxiety starts around six months and peaks at eight to nine. The eight-to-ten-month regression lands there — one to two weeks of it, for something like forty to sixty per cent of babies. New motor skills, rolling and sitting and pulling to stand, wreck sleep for days until they are mastered, and that is exactly the season for them. The third nap is going. All of that is in the books, separately. Nobody assembles it into the sentence a parent of a four-month-old actually needs, which is that the most broken nights are probably still ahead.
Why publish the bleak version
Because a night that is expected to be bad is a different experience from a night that feels like evidence you did something wrong. That is the whole argument. The information doesn't change a single thing about the night itself; it changes what the night means, and at three in the morning that is not a small trade.
The industry's instinct here is the opposite — everything is a phase you are about to defeat, ideally with the help of a purchase. It has to be, because a curve with a predictable hump in it sells nothing and a solvable problem sells a subscription. I would rather hand a parent the shape and let them plan around it.
The encouraging half is in the same numbers
As wakings got more frequent they also got shorter. Median time awake per waking: 27 minutes through the first six months, 20 at seven to eleven, 14 in the second year. Nights with no waking at all went none, 2%, 7%, 18%.
So the night doesn't resolve by the wakings stopping. It resolves by them getting cheap. That is a better thing to tell a parent at eight months than "it gets better", because it is specific enough to be checked — and, unlike "it gets better", it stayed true on the nights when it didn't feel like it.
One child, hand-logged, and a waking we slept through is a waking we didn't record. Same caveat and same direction as the measurement problem in the entry alongside this one: if anything these counts are floors.