I recently found myself having a completely erratic sleep schedule. I would go to bed at 9 P.M. one night, 1 A.M. the next, and every hour in between. To address this, I thought of beeminding when you go to sleep in terms of hours before midnight(or whatever time you want to set as your latest acceptable sleep time).
Say you want to go to bed around 9 P.M. on average every day. That’s 3 hours before midnight. I would begin by making a do-more goal with a rate of 2.5 hours a day. That way, I could build some safety buffer for those nights where I’m just stuck out late. Then I just enter when I go to sleep every day in terms of how long until 12. For example, going to bed at 10:15 would be 1.75 hours till midnight.
Let me know what you think of this metric, and alternative ways of beeminding consistent sleep.
I have used something to this effect in the past, it was quite useful at that period of my life to motivate myself to go to bed earlier.
What I did is I set IFTTT to post a “24” datapoint every day and then I manually entered the actual time when I go to bed with a negative sign, e.g. -21:30, which resulted in the sum of the two to be 2.5 hours. But obviously you can still manually enter the 2.5 hours.
(I am trying to remember if I did that to make use of some obscure aggregation method.
UPD: I looked into the archived goals and I am totally confused now. What I actually did is I had a do-less goal, with aggregation method set to “min”, and with a “24” datapoint posted automatically and an actual datapoint with a positive sign posted manually (or I tried some button-pressing trigger with IFTTT at some point as well). The point of “24” and “minimum” is not to gain buffer by not entering the datapoint at all, because this a do-less goal. But I have no idea anymore why do it so complicated; I would do what I described above initially if I did such a goal now. A do-more goal is much easier to wield, and you don’t need special precautions not to gain too much buffer by forgetting to enter.)
That’s a really neat way to solve a common problem! I guess you’re getting up around the same time, so the important thing is when you go to bed, and this solves for that nicely!
Very interesting, I like the idea of beeminding hours before midnight to allow it to be a do-more goal!
I’m beeminding going to bed earlier using a method that @dreev posted about ages ago—you can enter datapoints using the actual time, and Beeminder converts it to a fraction (for example, entering 9:30 gives you a datapoint of 9.5). It’s a do-less goal set to 10 per day and has definitely been helpful getting me to bed earlier. Also, no math, which is important to me!
I’ll just mention an alternative that I really liked using some years ago. Maybe it can inspire someone who is thinking about Beeminding their bedtime. I called it “bedtime rules”.
I basically decided on a time, say 22:30. When the clock hit this time, the rule was that I had to stop whatever I was doing, go to bed, and go to sleep without delay. I did not have to rush and I was allowed to follow my usual go-to-bed routine.
If I was in the middle of something that shouldn’t be ended abruptly, I would generally allow myself to do that within reason. For instance, I like to play competitive video games in the evening. A match usually lasts 10-20 minutes, and there is a penalty for repeatedly leaving a match before it concludes. I would allow myself to finish the current match and say thanks for the games/goodnight to my group. But I should not drag it out. If I was reading a book or watching TV, I would stop it immediately.
I don’t remember exactly how I did with being out and with people. I think that the rules were suspended as long as I was in an in-person social situation. But as soon as I left that social situation, the rules would apply again and I would have to seek towards my bed and go to sleep without delay. So doing anything other than going home and going to sleep after leaving the company would be breaking the rules.
If I judged that I had been following the spirit of the rules, I would give myself a point on a Beeminder goal the next day. The y-axis label for the graph was “cumulative total days of honoring the bedtime rules”. I can see that had a slope of 6 per week.
I think the subjectivity could be a problem for some, though it worked really well for me. I don’t recall ever being in doubt about whether I deserved a point or not.
The advantage was that it was much closer to an input goal. If I took any reasonable steps towards getting to bed, then I would be following the spirit of the rules no matter how late it became or what happened that prevented me from going to bed. The train is delayed for an hour? No problem. I can’t make the train go faster and I am still trying my best to get to my bed within reason. My video game match is dragging on, going to the 6th round? That is also not a problem, as long as I say my goodbyes and close the game as soon as the match ends.