Well, as expected the holidays sunk my wonderful trendline (see previous Nate Silver posts for context, otherwise this is lots of discussion about my battle with waistline). Anticipating that multiple xmas meals would undermine my weight loss plan I tried to max out exercise (got 6400 calories burned in 5 days) but even there I actually only managed to do 5 days instead of the usual 6 and actually fell a bit lower than last couple of weeks. So here’s where it shows up:
Once again the variability in the scale readings has a confusing effect, but given the minimum and maximum values of week 11 exceed week 10, the “trend” showing of 0.2lb gain seems about right (an interestingly is exactly the same as the difference in my “consensus” formula (((min+max)/2)+average+median)/3). So the effect is real, a gain of 0.2lb (worse than I hoped, better than my first measurement). That the range is so much greater in week 11 than week 10 is a little spooky but it’s hard to get a lot of data for these, plus I’ve now left home and eaten and …, no longer comparable conditions.
So, Nate, would you agree here with what I did?
Now my question is, how bad was this week compared to what I’ll need for maintenance. Despite the holiday I did show some restraint. Only 5 of the 7 days were off plan. And the meal part of those days was still restrained (it’s going to be the sweets and wine that kill me).
But this is all hard to reconcile. Due to the high variability of scale readings daily weights are not meaningful, so I can’t really compare today with yesterday, thus accounting for food consumption yesterday (zero exercise). So whether the last 24 hours was gain or loss (or how much) is unknown, but let’s look at it a different way.
According to an online calculator I should have consumed only 2228 calories for weight maintenance (given I did zero exercise, or at least relatively little above sedentary level). I had no “meals” at all in last 24 hours. So I consumed:
- around 3oz of green bean casserole (had butter and cream), so let’s say, 450 (probably a little high)
- I estimate I had about 7oz of my own bread (part of the “snack” I had for watching football, in lieu of dinner meal), which seems to work out to about 586 calories
- Now the killer, brie (with the bread), but both the quantity and calories are smaller than one might expect (I weighed what was left and deduced what was consumed from weight on package, and divided by two since it was two of us eating this “snack” at about equal frequency (I might even have been a little restrained)), so that comes in as 3.33oz and 320 calories
- And the even worse (or at least I thought) killer, the wine, but even though I indulged (less than my old days), that probably hits for 400 calories (so worse than the brie, but less than the bread). So my total “snack” was 1306 which pushes the worst I could get with a fast food meal and was probably about 700 calories more than what I would have otherwise eaten as a “meal” as part of my diet.
- Now even though I went geodashing I resisted all the snacks I would have normally eaten and essentially had zero calories during all that, so really I’m only up to 1756 that I can account for.
- But there are two other items: popcorn (naturally popped with oil, not the worthless air-popped which is worse than eating cotton swabs and therefore why bother) and almonds. The almonds have been a bad slide for me. Lots of “nutrition” books tout the value of almonds so they seem like harmless snacks. But the package claims 180 cals for 1/4 cup and the USDA says 169/oz (let’s see, Net sources says 1 cup whole almonds weighs 168g, so 1/4 cup is 42g which is 1.48oz, so USDA is saying 250cals/0.25cup, is the package lying?) Now I don’t have any records nor accurate memory of this (obviously this is a new area to focus on for control) but I’ll guess about 3/4 cup and splitting the difference on calorie estimates, that would be about 650 calories, and that is possibly the killer.
- Estimating the popcorn is tough. The USDA site has value for the popped volume and from another site I learned that 1/4 cup unpopped equals about 4 cups popped. Now I don’t know the precise amount but I think it was just over 1/4 cup, so let’s see 5 cups popped and I ate 3/5th so that’s 3 cups for me and USDA says that will be 420 calories (note that is worse than the brie), so now I’m up to 2826 (and I’ve tended to guess high on each individual element so I suspect this is close to worst case).
- So this translates to around 600 calories too much yesterday, which is almost my gain for the week.
But that doesn’t make intuitive sense. Yesterday was bad, but it wasn’t my worst day. I think at least two of the days were worse, maybe three, and only one was a lot better, so I’d estimate yesterday at a little below average and therefore my “excess” I computed here is too much. OTOH, this might actually show the exercise effect. Other days were worse from food consumption POV but much better from exercise POV (like 1300 calories more). So that 1300 calories buys me quite a margin for extra eating, so it might be that yesterday was nearly the worst (not so bad on excess food, but way worse of exercise to burn off the food). So maybe yesterday is the explanation. I think I had one good day, which based on past patterns would be (2.7lbs/7)*3600=1400 cal/day deficit (which, btw, tends to mean about 850 due to exercise and so only 550 due to reduced food intake, meaning I’m hitting about 1700 cals/day eating). So, for the week: four days at zero (extra intake offset by exercise), one day at -1400, yesterday at +600, to explain the 0.2lb gain). IOW, if I’d resisted the football snacking yesterday I might have finished the week flat (which was my worst case goal). IOW, it doesn’t take much of a lapse to offset all the effort I made on the other 6 days (still had some restraint in holiday meals, plus lots of “excess” exercise). So the vigilance that maintenance is going to take is going to be staggering. And trying to do calculations has too much error and doing weigh-ins has too much error, so compensating for “bad days” (“slips”) but a reduced day and extra exercise is going to be difficult.
But now what connection does any of this have to Nate Silver. Well, here’s the problem, my previous “money graph” is now blown up:
The trendline on the left if first 10 weeks and the trendline on the right includes this weeks definitely anomalous point. So what should I do for future graphs/analysis (options: a) just keep adding points, b) break into two line segments starting on week 11 and then future weeks). Assuming week 11 was a total anomaly not to be repeated and the previous trend will resume then two segments would seem best (as a predictor). But that’s going to be a bit complicated because: a) this week is not going to be “back to plan” because football games and more holidays are bound to keep me away from my discipline, and, b) the pure linear weight loss was doomed to stop anyway (the issue was that as weight was dropping, two counter-tendencies occurred: 1) exercise burned less calories at same amount of exertion, 2) basal metabolism requirements drop). So the 2.7lbs/week (the slope of trendline up through week 11) was not sustainable. I would “predict” (pure seat-of-the-pants, no Nate Silver methodology) now an achievable level of only 2lbs/week for the next 25lbs (and even less after that). So maybe next week, where it’s unlikely I’ll achieve 2.7 is just the start, so again two segments will be useful.
But what do I want out of all this fanaticism of messing with numbers and graphs – well, as Nate wants, a prediction. Accumulating data is slow and I want results ASAP so I don’t want to only be able to make some prediction (like when I reach 200lbs) by the end of next month, so I need to use the earliest data (and only a couple of weeks of it) to make an accurate prediction. So, as I’ve said before:
need some help here, Nate!
So I’ll make a prediction to check next week. Even though this coming week won’t be as good as my plan I’ll still shoot for:
and thus all the way out to 20Jan2013 to finally record weight below 220lbs which was approximately my starting point when I started training for triathlon (27 years ago) and reached just under 200lbs. This also means I’d still be on track to crack 200lbs by tax day (with a little to spare, I think, if I can stay disciplined). So if I can then stay disciplined and do slightly better than maintenance I might be “fit” (belly gone) by birthday next year so that seems to be the goal I should have.