Once upon a time I gained a lot of weight (20+kg, crossing the "obese" BMI bar) over a dozen or so years, and started to physically feel really bad. I realised

Once upon a time I gained a lot of weight (20+kg, crossing the "obese" BMI bar) over a dozen or so years, and started to physically feel really bad. I realised that if I want to save my life (as in not dying from a heart attack or something but also being a living entity while not feeling absolute garbage every day, which would be wasting my life) I had to do something. Since it took me that much time to gain that weight and wanted to go back to feeling my best in a way that sticks I set up a weight loss plan that would unfold over the same amount of time than it took, because there's just no shortcuts to it. So I did a big ass spreadsheet, plotting rolling 7 day + rolling 24 days and their derivatives. Key insights of my strategy: - instant measurement does not matter, as in it's actually not worth looking at every day, only retrospectively when observed as a collection upon looking for root causes of deviations of the averages. most advices about weighting in at a certain time or a certain day of the week is about faking an average with discrete measurements (which kinda works but requires extremely effortful consistency to make sense, and any error throws the fake average off big time) - 24 day average is really nice and meaningful: missing or mismeasuring (not weighting early morning+ after poop, scale glitch...) a few measurements would be inconsequential, a couple stray events would not affect it, a positive derivative would mean I must take immediate action (as in, get back on track with being consistent) to correct course and reverse the trend. when in the losing weight phase, a negative derivative closing in towards zero would mean I should pay a bit more attention to being consistent. as target weight† nears or is attained the meaning of that derivative shifts a bit (obviously one does not want to perpetually lose weight, so things offset and taper off) - if weighting multiple times a day for averages it would consider only the minimum for that day, but it would also give me a spread for a given day, allowing me to understand fine grained details of my body behaviour in face of specific events (overeating, undereating, doing sports, drinking water, drinking alcohol when partying) and compare that to the next day(s). such events preceding each measurement were annotated on the corresponding row. - from an initial data gathering and retrospective, 24 day appeared to be my body latency: when I did something effective or veered off course the effect would visibly settle in as long term weight loss or gain after that amount of time (this would be visible in raw data, not artificially come from the average construction, so I tuned that average to 24 days to produce the corresponding synthetic, immediately meaningful derivative value) - 7 day allows for shorter term observation and reactions, but is more susceptible to noise. It appeared from observational data that it was the rough scale at which things would start to have long term effects happening at t+latency. e.g not doing sports, or overeating, or bad sleep from t to t+5~7 would produce a long term, lasting effet at t+20~25 days; lower than that and it would just evaporate. overall it removed a whole lot of complexity, guesswork, and guilt, which translated in not skipping measurements, which produced more and better data, making the whole process a non-event, progressively building healthy habits, making it easier to course correct, and so on and so forth. the end: I shed all the excess weight two years before the "deadline" without feeling that I put much effort into it (there was effort, especially when I needed to kick my ass a couple times, but it did not feel like I was constantly abusing my body; the "consistency" effort was spread out and thus much more manageable). concretely I ended up both feeling (internal assessment) and looking (external feedback) younger in my late 30s than in my late 20s. I've since replaced that with a Garmin watch + scale and the Garmin app does pretty much what my spreadsheet did, minus the derivative bits, which I really miss. † I never aimed for an actual weight value, but for an intrinsic feeling of healthiness, which for me happens to correspond to a specific weight range.