Pushing the Optimization Further

Adding more sleep to support a higher training load.

I wrote an entire article explaining why I had reduced my sleep to five and a half hours, and I am now increasing it again.

This is not me abandoning the experiment. Annoyingly, it is probably the logical conclusion of it.

The original schedule was mostly designed around cognitive work. Four and a half hours of core sleep, three twenty-minute naps, and a day split into shorter blocks so I never got to the completely useless tail end of a normal waking day.

Then I started increasing my training.

As a calendar problem, the old schedule was great. I had more hours.

As a recovery problem, it started making less sense.

Exercise does not just need time. It needs sleep too.

So the change is pretty simple:

• Old: 4h 30m core + 3 × 20m naps = 5h 30m
• New: 6h core + 3 × 20m naps = 7h

The nap structure stays. The core sleep just gets longer.

The only small timing change is the final nap moving from 17:00 to 17:40. Everything else is basically the same idea with a larger recovery budget.

Why increase it now?

Because I want to train more.

Not in the vague sense of "be more active." I mean more running, more strength work, and enough consistency for it to actually compound.

The mistake was treating that as a scheduling problem first.

If I want more exercise, I can obviously make more time for exercise. Cut something else, wake earlier, move blocks around, done. The calendar is not the hard part.

The harder part is whether the workload is recoverable.

I had compressed sleep to create more usable hours, then started increasing one of the things that makes sleep more valuable. Once I wrote that down, the contradiction became difficult to ignore.

This is the same recovery-budget idea from the first article, just pushed one step further.

A harder run, a strength session, poor sleep, and a long day do not exist independently. They all change what I can reasonably ask from the next block, and from tomorrow.

So I am giving the system more recovery before I give it more load.

The part I am keeping

I still like the nap architecture.

That was the useful part of the original experiment. The naps break the day into shorter waking blocks and give me hard boundaries instead of one long stretch where fatigue quietly accumulates in the background.

I also still like the idea of managing state instead of just managing hours.

The change here is not philosophical. It is almost boringly practical.

Five and a half hours was aggressive. It becomes even harder to justify when I am deliberately increasing training volume.

So I am moving the core block from 22:00 → 02:30 to 22:00 → 04:00.

That is it.

What I am testing now

The question is whether I keep the parts I liked from the old schedule while getting enough recovery to support more training.

I am watching the same things as before: decision quality, irritability, whether the naps still happen naturally, whether I make stupid mistakes late in the day, and whether the schedule survives contact with real life.

Now I am adding training performance and recovery to that list.

If seven hours still turns out to be too little, I will increase it again.

There is no point defending a number because I happened to publish it once.

The first version was an experiment. This is the next version.

The original article is here.