Sleep was the variable I kept ignoring

I logged intake, weight and exercise, and treated sleep as background - something that happened rather than something that participated. Going back through several months of records, the pattern is embarrassingly clear.

What the log showed

The weeks where I ate noticeably over budget were, with very few exceptions, the weeks following a run of short nights. Not the stressful weeks and not the busy weeks specifically - the under-slept ones.

The mechanism is well documented and I will not pretend to expertise on it: appetite regulation is affected by short sleep, and the appetite in question is specifically for quick calories. What I can report is that on the days after five hours, my own estimates of what I had eaten were also the least accurate. Tired judgement is generous judgement.

What I changed

• A fixed wake time, including at weekends. Easier to hold than a fixed bedtime, and it drags the bedtime along behind it.
• No screen in bed, which I resisted for a long time and which made more difference than anything else on this list.
• Last coffee before two in the afternoon.
• Logging sleep hours next to the weight, so it stopped being invisible.

None of it is novel and none of it is hard. It just had to become something I measured before it became something I acted on - which is the same lesson as everything else here.

As above: this is one person's record, not medical guidance. Persistent trouble sleeping is worth raising with a doctor rather than solving with a blog post.