So in light of feeling the brunt of all these attempts at habit formation, I’ve stepped back and tried to look at a better plan based on all the stuff I’ve been reading.

So here’s my theoretical framework: If self discipline is one deplete-able resource, and if automaticity is when an action approaches an asymptote of 0 willpower, then the best way is to have a slowly unfolding progression of new actions taking over as the first action slides towards that asymptote.

For example, if drinking a glass of water before breakfast takes 18 days to form into a habit, then I’d want to start another habit at some point past the mid point where the slide to full automaticity is underway.

I’m assuming, perhaps erroneously, that tasks take a certain amount of willpower on day 1, and start getting harder as the “streak” of sustaining the habit continues as days go by. This certainly is true for my own personal attempts. Going to the gym is easy on day 1 or 2, but it gets more and more difficult to continue the streak of days (obviously there are some random days where it’s more or less difficult).

Given the above, and given that different tasks take different amounts of times to reach automaticity, then it makes sense to start Habit 1 until the midpoint of automaticity, then start Habit 2 until it reaches ITS UNIQUE midpoint, etc. This way the process of automaticity itself shoulders the brunt of sustained will - rather than how I’ve been doing it now - which is simply adding tasks 1 week later regardless of their difficulty or willpower expenditure.

So now the question is, how do you establish midpoints? I do not want to test each task, so what I’ll do is ball park it using Lally’s experiment - I’m assuming that tasks that take the least willpower will take the quickest to form into habits and the ones that take more will take longer.

So with Lally’s data, I can estimate that easier tasks will take 18 days, middle of the range will take 66 days, and the most difficult will take 250 days - I’ll divide each to peg the midpoint, and estimate the variations between tasks that I’ve already labeled as Easy, Medium, and Difficult.