How much do I weigh?

October 11, 2015 1:36 PM

How much do I really weigh? A google doc spreadsheet and a screenshot of Excel's graph. Most weighings are on the scale at the gym after riding in to work.

My low of 200.8 after cycling 40 miles in 90 degree heat on Sunday, then 11 miles to work on Monday morning. The big drop from the 210's to 205 from July 1 to 29 a) was after riding 20-24 miles per day, 4 days/week through 4 weeks in July b) tells me I should weigh less frequently to see any real gains.

I'm 6' 3 1/2", 48 years old, male and I don't get much exercise besides riding and walking. I should probably add core & arm stuff to get down to my target 195-200 and appear more "balanced." My high was 243 lb. a few years ago (2011, I think).

My main question, though, is whether this 2-3% variation is unusual. I'd love to hit < 200 lb., but if I'm 207 the next day, it's more frustrating than motivating. I'm using the gym's scale because after finding big discrepancies with my home one, I brought it to the gym & it was reading about 3 1/2 lb. higher.

posted by morganw to Health & Fitness (9 answers total)

When I was really into losing as much weight as possible, I calculated a five-day average. That's a much better way of tracking weight loss. I weighed myself every day and computed a five-day average in Excel, plotting this on a graph overlaid with my daily losses. It showed a clear regression line of steady loss that was very motivating.
posted by sockermom at 1:43 PM on October 11, 2015

2-3% variation is completely normal. If you weigh in daily, the likelihood of weighing in under 200 one day and then never weighing over 200 again is almost nil.
posted by telegraph at 1:54 PM on October 11, 2015

Yeah, variation is very normal. There are a few other factors which can affect it, but mostly the short-term variation is dictated by hydration level --- remember that you're mostly water, and that the exact amount of water in you changes constantly. Significant weight change is a long-term process as a rule; keep daily records, but look at trends rather than fluctuations.
posted by jackbishop at 2:21 PM on October 11, 2015

If I were you, I'd read John Walker's free ebook The Hacker's Diet: specifically, the chapters on "The Rubber Bag" and "Signal and Noise." Walker suggests using an exponentially smoothed moving average to figure out your weight trend.

I weigh myself every morning in my birthday suit, after first visiting the toilet. My weight has fluctuated by as much as 5 pounds from day to day, or over 3% of my total body weight. The average difference in my weight from day to day since I stopped losing weight and have aimed to maintain is 1.00 lb.: in other words, fluctuation is normal, and seeing the same number for more than two days in a row is kind of freaky.

Over the course of a day, my weight can vary even more, since I do long distance cycling events from time to time, where a little dehydration is impossible to avoid.
posted by brianogilvie at 3:14 PM on October 11, 2015

I work for a scale company and have daily access to very accurate scales that are accurate to 0.01 lb. Yes I can stand on them and yes, I can weigh my poops by the before/after difference. In the biz we call that the gross, tare, and net.

A few pounds variation is completely normal. You will lose a couple of pounds at night exhaling. Seriously. You breathe in O2, and exhale CO2, which is heavier by a carbon atom but takes the same amount of space. Those carbon atoms came from your food. You lose them by breathing, all the time.

You also have a huge dynamic (at least on this scale) by drinking and perspiring. You probably don't think of fluid as "weight" because it's not food, but water that's in your body tips the scale. This can be a bit pernicious because dehydration makes the scale go down but it's actually a bad thing. If you want to lose weight what you want is for your hydrated weight to be low, meaning your body has discarded actual tissue mass. Your body can absorb and release several pounds of water in the course of a few hours.

Which is by way of saying, it's usually unproductive to focus on gains and losses of a pound or two on short time scales. Our bodies are very dynamic and there are tricks (some exploited by "diet" pills) that can cause surprisingly extreme short-term changes. The meaningful question isn't what you weigh right now to 0.01 lb, but what the trend is over a period of at least a few days.
posted by Bringer Tom at 6:38 PM on October 11, 2015

I have found a wifi scale to be a stupidly expensive but really helpful investment for this. I can step on the scale every day but don't have to look at it, the numbers just disappear into the cloud to be recorded. Then I can look at my rolling average every few days and get a smooth estimate of where I'm going.

(And read the directions on your scale--use a smooth, flat, non-carpeted surface. Zero the scale fairly regularly. Weigh at the same time, with the same circumstances, every day. It's more about consistency than the exact number. It's OK to use a scale that is 3.5 pounds "high", as long as it's consistently so--getting the number 195.0 to show up will not bring happiness. The goal isn't a number, it's a specific body composition, and the scale is a tool to help you figure out if you're making progress toward the goal.)
posted by anaelith at 5:51 AM on October 12, 2015

So what you could do is weigh yourself on 30 consecutive days without trying to lose weight just to get a baseline. Then you could calculate a mean and a standard deviation. So then you start trying to lose weight, with the null hypothesis of "I have not lost any weight." Every measurement, you could calculate how many standard deviations from the mean that is and convert that into a probability that that measurement would occur if the null hypothesis is true. A couple of pounds difference is going to be pretty likely, but eventually you're going to reach a point where your measurement is so unlikely that you have to reject the null hypothesis and conclude that you HAVE lost weight.
posted by ctmf at 2:59 PM on October 12, 2015

That makes a bunch of probably-not-valid statistical assumptions, but it would be fun anyway. If someone asked you if you were losing weight, you could tell them numerically how confident you are that you are.
posted by ctmf at 3:02 PM on October 12, 2015

I will second the recommendation of the Hacker's Diet website, not necessarily because of the diet or because you should read any of it, but because it has an online weight tracker that calculates that rolling average that you should be using.
posted by CathyG at 9:09 PM on October 12, 2015