Tuesday, March 15, 2011
OPT
Sort of following up on this entry about frugality and retirement.
Over the last few years I've shifted rather quickly toward Old People's Time. For some reason this artifact of aging surprised me; it's well known but I hadn't observed it in parents or grandparents.
My default sleeping pattern (Young Peoples Time or YPT) was fairly constant from age 8 to 58. Hit bed between 1 and 2 AM, wake up around 9 AM when not forced out of bed by horrible emergency situations such as work or school. This pattern was broken by a few years of night-shift work in the '70s, but returned to YPT default when I returned to day shift.
OPT started to creep in a few years ago, and now it's fully locked in. Bedtime is 8 PM invariant; wakeup is between 2:00 and 3:30 AM, with optimum at 2:45. That's a big change of habit in a few years! 2:45 was the tail end of bedtime range, now it's the middle of wakeup range.
To pin down the pattern I decided to extract times from the HTML archives of this blog, because I've always tended to post either at the exact start of the day or near the end. Often it's the need to write up a "new" idea that pulls me out of bed.
Here's the result after some Python processing and Excel spreadsheeting.
Horizontal axis is months, vertical axis is time in 24-hour format.
Bottom line (blue) is earliest post time for each month, which equals wakeup time; top (violet) is latest post time for each month, which is generally an hour before bed; middle (yellow) is mean of all dates in the month, which consistently favors the early side. On days when I write more than once, the second writing is usually in the morning.
The zero outliers on the bottom line are after midnight, which really should have been counted as latest instead of earliest .... but I couldn't think of a way to flip them without a lot of manual searching through HTML.
Overall the trend is clear, and of course it's a tanh or sigmoid, like most patterns in nature. Looks like OPT began to take effect in fall 2008, which makes some sense. Two serious strains happened in that season. First the Grand Theft Of America by Shotgun Paulson, which was a real mental shock. (I already knew the system was rigged, but Paulson forced the conclusion that the whole ruling setup is pure Mafia.) Then I got hit by a car, which was probably enabled by the cognitive shock.
= = = = =
Later and broader comment on OPT here. = = = = = Update April 2015. Added later years to the XLS and graph, making a 10-year record. Also reprocessed to make more sense of times between midnight to 2 AM. Before 2011 those times were at the end of my day; after 2011 they were at the beginning. (Had to treat 2AM as 2600 hours for the end-of-day instances, which is why Excel runs the graph up to 3000.) This gives a much stronger and clearer pattern to the whole thing:
Not hard to tell when I decided to switch bedtime!
¶ 9:41 AM 
Over the last few years I've shifted rather quickly toward Old People's Time. For some reason this artifact of aging surprised me; it's well known but I hadn't observed it in parents or grandparents.
My default sleeping pattern (Young Peoples Time or YPT) was fairly constant from age 8 to 58. Hit bed between 1 and 2 AM, wake up around 9 AM when not forced out of bed by horrible emergency situations such as work or school. This pattern was broken by a few years of night-shift work in the '70s, but returned to YPT default when I returned to day shift.
OPT started to creep in a few years ago, and now it's fully locked in. Bedtime is 8 PM invariant; wakeup is between 2:00 and 3:30 AM, with optimum at 2:45. That's a big change of habit in a few years! 2:45 was the tail end of bedtime range, now it's the middle of wakeup range.
To pin down the pattern I decided to extract times from the HTML archives of this blog, because I've always tended to post either at the exact start of the day or near the end. Often it's the need to write up a "new" idea that pulls me out of bed.
Here's the result after some Python processing and Excel spreadsheeting.
Horizontal axis is months, vertical axis is time in 24-hour format.
Bottom line (blue) is earliest post time for each month, which equals wakeup time; top (violet) is latest post time for each month, which is generally an hour before bed; middle (yellow) is mean of all dates in the month, which consistently favors the early side. On days when I write more than once, the second writing is usually in the morning.
The zero outliers on the bottom line are after midnight, which really should have been counted as latest instead of earliest .... but I couldn't think of a way to flip them without a lot of manual searching through HTML.
Overall the trend is clear, and of course it's a tanh or sigmoid, like most patterns in nature. Looks like OPT began to take effect in fall 2008, which makes some sense. Two serious strains happened in that season. First the Grand Theft Of America by Shotgun Paulson, which was a real mental shock. (I already knew the system was rigged, but Paulson forced the conclusion that the whole ruling setup is pure Mafia.) Then I got hit by a car, which was probably enabled by the cognitive shock.
= = = = =
Later and broader comment on OPT here. = = = = = Update April 2015. Added later years to the XLS and graph, making a 10-year record. Also reprocessed to make more sense of times between midnight to 2 AM. Before 2011 those times were at the end of my day; after 2011 they were at the beginning. (Had to treat 2AM as 2600 hours for the end-of-day instances, which is why Excel runs the graph up to 3000.) This gives a much stronger and clearer pattern to the whole thing:
Not hard to tell when I decided to switch bedtime!
¶ 9:41 AM
