Most self-experimentation is one person, unblinded, taking a thing and deciding it worked. That design cannot separate the compound from the expectation, the season, or the fact that you started sleeping better the same week. This turns the same experiment into an actual trial: randomised order, self-blinding, a washout sized from the half-life, and an analysis you commit to before you have seen a single number.
Nothing is uploaded. The schedule is generated in this tab from a seed you can write down.
The thing you take, with dose and timing. "Magnesium" is not a protocol.
One primary outcome. Everything else you record is secondary and cannot rescue a null.
Sets the washout. Unknown? Use the longest plausible value.
Each pair is one on-period and one off-period, order randomised.
Write this down. It is what proves you did not re-roll until you liked the order.