Applied Nate Silver – a new way to monitor weight

I’ve now finished 70 weeks of my close monitoring of my weight, trying to use good statistical techniques to find the noise buried in the signal. To refresh this idea I face two major noise issues in my data: 1) the high variability in my digital scale, and, 2) my daily fluctuations, which I still can’t pin down. And again to review history, basically I had a nice steady 24 week period of 2.4lbs/week loss that got me close to my target weight (BMI just under the overweight category, for what BMI really means), then a few disruptions due to vacations, then a period of near steady state approximately at my target, then falling off the wagon due to medical issues, and then my attempts to get back down to steady state.

What I’ve found is that using statistics and graphs (as incentive and/or discipline) works better while losing weight to reach a target than maintaining. In steady state I don’t get the “reward” of a nice declining value and so it’s hard to continue starving myself on a daily basis. And despite averaging over two hours/day and 1200 calories of exercise I do still have to be hungry all the time (forget silly fad diets – the only answer, at least for my metabolism, is hunger).

And so recently using Sunday as my weigh-in day has fallen apart. Sunday is a bad day to pick since Friday and Saturday tend to be “fun” (i.e. eating days) and thus the drop I have during the weekdays falls apart by Sunday (and worse, on Sunday) as shown in recent data below:


The pattern, outlined by the yellow lines, shows my tendency to fall off the wagon and bloat back up again, then followed by renewed resolve. Note that Saturday I hit a new (recent) low but, bam, back up again.

So Sunday (and also a single day) may not be the best way to measure this and also motivate myself so here’s another approach:


Here I’ve taken the individual scale values (5 per day on weekdays, 7 on Sunday) as represented by the small blue markers, all consolidated into a single interval (all seven days of the week combined). Then I get the median (red circles), low (green triangles) and highs (squares) and do regression lines for each (approximately the same slope). While these values bounce around too what this shows is my records on weekly, rather than a single day. So maybe I can use these as my feedback. If I can’t maintain fasting discipline for a day or two, at least maybe I can do it for the entire week.

Note that the weekly change is running about 1.3lbs/week (I’ve now been doing that for 12 weeks), but here’s how severe the noise is (both scale and daily body fluctuations) are:


The bars represent the difference between highest and lowest measured weight during the week and the line represents the average of these (6.3lbs). IOW, the noise (6.3lbs) is way larger than the signal (1.3lbs). On one hand seeing the extremes is a good motivation to me but I feel too guilty (after a bad uptick) and that’s not the best motivator (does tell me to renew my resolve to control, i.e. starve as best I can). But it’s so easy to have a one-day fluctuation (or worse, just scale variability) compared to my longer-term progress than it’s very demotivating. So, on one day, a bad number shocks me into action, but then knowing I’ll blow the weighin the next day I just say, to hell with it, eat as I wish.

Now the food scolds would say I could solve all this with kale and quinoa and oatmeal, but that’s silly. Whatever causes these daily body fluctuations is very difficult to control. And I like to cook tasty food every now and then, plus I live with someone who wants real meals, not reduced calorie starving. And as we’re getting glimmers of better weather there is pressure to grill outside and it’s impossible to not have upticks whenever I cook good food (btw, food nags – I rarely eat junk food, it’s “real” food (and unfortunately some wine) that causes the upticks; only the artificial calorie limited meals work to get downticks).

But one thing I’ve learned is not to fool myself by simply cooking the books. When one weighin per week was my “money” graphic the pressure of that weekly deadline helped with resolve. If I start blurring my statistical measurement with lots of values then I can do the usual rationalization (eat today, starve tomorrow to make up for it, but it’s always much easier to do the eating than the starving).

So I do need some flexibility, especially while at my target weight and just trying to maintain, not lose, but not so much that I make too many excuses for myself.

The food nags think this is simple, i.e. just do some stupid fad diet or vegan nonsense, and everything will be fine. The only part of their logic that makes sense for a real human being is glycemic control when you really need to do that (one of their stupider ideas is that slowly absorption of calories is useful, not for weight control it isn’t, calories are calories whether my intestines absorb them slowly or quickly). So all these inherently skinny people who act so damned superior really have no idea what it’s like when your natural setpoint is too high.

Human beings evolved in an environment of relative food scarcity, no pantry to just go get more. So natural selection favors creatures who can pack on a few extra pounds in order to survive when food really is not available, not just a choice not to overeat. So I have those positive survival benefit genes – gobble food when it’s available, starve the rest of the time when it’s not. Of course, unless I get too poor calories are easy to come by in my environment so being hungry has to be by choice, not necessity. And for me I can’t just do some stupid fad diet to achieve that – I need this statistical discipline to then give me the resolve, hour-by-hour, day-by-day, to continue starving myself most of the time.

The only “diet” that makes any sense is eat less (no matter what, there is no magical superfood silver bullet to help with this, I’d get just as fat as a vegetarian than an omnivore)) and move more. Moving more is the easy part (at least until my arthritis gets worse) so burning a 1000 calories/day in exercise is the easy part for me. It’s the constantly being hungry that is the hard part (and sorry, olive oil, nuts, and yogurt is just as fattening as junk food, it’s 100% a matter of caloric intake, not what sources).

So hopefully I can find a new statistical approach that will work.


About dmill96

old fat (but now getting trim and fit) guy, who used to create software in Silicon Valley (almost before it was called that), who used to go backpacking and bicycling and cross-country skiing and now geodashes, drives AWD in Wyoming, takes pictures, and writes long blog posts and does xizquvjyk.
This entry was posted in comment and tagged , . Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s