Test one thing at a time. It feels slower and it is the only way to learn anything.

Ten percent cuts, of the current dose not the original. That is the arithmetic that keeps it survivable.

An inconclusive account is useful if it stops the next person mistaking a story for a controlled comparison.

Three questions kill most of the confidence in a thread like this: sample size, blinding, who paid.

Log dose, time and sleep. Ten seconds a day and it turns the anecdote into something you can look at.

Test one thing at a time. It feels slower and it is the only way to learn anything.