Weight fluctuation in day

Silentpadna wrote: »

Christine_72 wrote: »

Because weighing in the morning before eating and drinking anything is about as accurate as you're going to get. Why would i weigh myself at night when there's a full day's worth of food and fluids sitting in my gut, which will give me a skewed and rather depressing number.
I put on a pound after drinking 2 cups of tea, I dread to think what a fulls day of eating and drinking would do to the scale number.

Why is your weighing time in the morning any more accurate than the evening over the long run?

That said, I do understand the idea of avoiding the "depressing number", and the idea that positive feedback matters. I think all of us on this weight fat loss journey would do well to remember that the scale is a tool, but it doesn't can't measure fat loss in the short term.

I got on this shtick in many threads because I see a lot of folks making decisions based on their feelings rather than facts. And I understand it's hard when the scale is, for most of us, the only tool we really have. It's hard to trust the process over the long term when most of us, me included, want quick results. For me, I have to continually choose to trust the process that we have seen proven time after time. Be as accurate with measuring input and output as you can and refine the process over the long term.

50 posts in and you think us saying to weigh at a consistent time with the fewest variables is to massages our fragile egos by seeing the lowest number possible? You don't know what you think you know.

It is often said by regulars here that psychologically, if you can deal with the fluctuating numbers, weighing daily is a great way to track trends over time (not 1 week but more like 8 weeks to see proper meaningful data) and take all the emotion out of the numbers. It is also recommended to use other metrics such as body measurements and fitness goals to track progress.

The people who come here freaking out over a week with no loss or a spike one day are always advised that one week or one weigh in is not long enough or enough data points to need to worry or start making drastic changes.

TL;DR Your point is being erroneously made and you're projecting feelings onto people that they don't have over the scale number.

stanmann571 wrote: »

Because, weighing in the evening, introduces variables like
Did I drink 3 or 5 liters of water today.
Did I have an extra bite of something with dinner.
was there extra electrolytes in the seasoning with dinner, etc

Weighing in the morning, after first BM, before first feed.
minimum of 6 hours dehydration/sleep/digestion/etc.

But in the long term, do those variables mean anything? That's my point. Those variables are there all the time, whether we are trying to lose, trying to gain, or trying to maintain. Who says those seasonings aren't having an effect in the morning? What if you're not blessed in being regular on your BM? Accuracy day-to-day doesn't really matter. Long-term, those variables go away when you're trying to determine whether you've lost fat. There is no real accuracy issue when you are looking at a big enough time window to make any decisions.

I see that I am having a very hard time getting my point across - and that means it's on me, not those who are responding to my admittedly contrarian ideas here.

We can all agree that fat loss is not short-term, right? We can also agree that fluid retention can mask fat loss (or even gain) in the short-term, right? In order to see if we are achieving fat loss (assuming all we are using is a scale), we have to give enough time for the fat loss to be substantially greater than day-to-day fluctuations. If we are losing 2 pounds per week in fat and we fluctuate by up to 8 in a day (or for some people who have mentioned even more), you need probably two months to realistically see fat loss and eliminate noise. If you are weighing yourself every day, what difference does it make what time it is if you cannot see the changes anyway?

If I wanted to eliminate as much noise as possible, I would reduce the number of times I weigh myself to at least a month, probably two - and in those cases where I'm actually doing it to record progress, I would then want to make the conditions consistent (even though that's probably also not very likely). Personally I look at the scale a number of random times during the week just to look at the data, but really all it does is confirm fluctuations. Entering random weigh-ins on trending apps yields much the same results. The trend looks fine.

Silentpadna wrote: »

stanmann571 wrote: »

Because, weighing in the evening, introduces variables like
Did I drink 3 or 5 liters of water today.
Did I have an extra bite of something with dinner.
was there extra electrolytes in the seasoning with dinner, etc

Weighing in the morning, after first BM, before first feed.
minimum of 6 hours dehydration/sleep/digestion/etc.

But in the long term, do those variables mean anything? That's my point. Those variables are there all the time, whether we are trying to lose, trying to gain, or trying to maintain. Who says those seasonings aren't having an effect in the morning? What if you're not blessed in being regular on your BM? Accuracy day-to-day doesn't really matter. Long-term, those variables go away when you're trying to determine whether you've lost fat. There is no real accuracy issue when you are looking at a big enough time window to make any decisions.

I see that I am having a very hard time getting my point across - and that means it's on me, not those who are responding to my admittedly contrarian ideas here.

We can all agree that fat loss is not short-term, right? We can also agree that fluid retention can mask fat loss (or even gain) in the short-term, right? In order to see if we are achieving fat loss (assuming all we are using is a scale), we have to give enough time for the fat loss to be substantially greater than day-to-day fluctuations. If we are losing 2 pounds per week in fat and we fluctuate by up to 8 in a day (or for some people who have mentioned even more), you need probably two months to realistically see fat loss and eliminate noise. If you are weighing yourself every day, what difference does it make what time it is if you cannot see the changes anyway?

If I wanted to eliminate as much noise as possible, I would reduce the number of times I weigh myself to at least a month, probably two - and in those cases where I'm actually doing it to record progress, I would then want to make the conditions consistent (even though that's probably also not very likely). Personally I look at the scale a number of random times during the week just to look at the data, but really all it does is confirm fluctuations. Entering random weigh-ins on trending apps yields much the same results. The trend looks fine.

IF you weigh less often, you become more vulnerable to noise.

The more data points you gather, the smoother becomes the curve of your data.

stanmann571 wrote: »

IF you weigh less often, you become more vulnerable to noise.

The more data points you gather, the smoother becomes the curve of your data.

Why is that? Doesn't it depend on the window of time you are taking data? What if you took 20 data points over a one week stretch, and I took 4 over an 8 week stretch. Which do you think would give you the most accurate trend? The one with more data points or the one with more time? Further, If I took 4 over an 8 week stretch and you took 40, would our trend lines be substantially different?

What is the smoothest curve? Either a linear function (no curve), or a perfectly functional curve where the the dependent variable (weight in this case) is precisely related to the independent variable (time). We know it's not, because weight itself is comprised of more than one thing and they change at different weights due to totally different circumstances. As a result, we end up with a trend line (or curve) that can be calculated using data points. If our goal is to lose fat we need a long time window to see the trend line because fat weight changes slowly.

Look at the data for stock prices or market indexes. Over the very long term the curve is much more smooth than shorter terms. Why? Because daily, weekly, or monthly fluctuations vary much more dramatically than long-term ones. If you are looking for what will happen to a stock price or index over a very long period of times, it doesn't matter at all what time of day you check its price on a given day. It's not a perfect mirror for weight fat loss, but it does illustrate that the more data points there are, the more noise there is. But...the number of data points don't affect the trend...they are simply components of it.

Silentpadna wrote: »

Tacklewasher wrote: »

And, by picking consistent conditions (first thing, after peeing, before eating) you eliminate as much of the noise as possible.

It's not perfect. But nothing will be. Even a daily Dexa scan would be imperfect. It would be better but is cost prohibitive.

Besides. Weighing daily can teach you something about how food impacts. So Wed was Sashimi for lunch and that means too much soy sauce. Thursday I was up 1.6 lbs (expected more). But no bad feelings about it and 1.3 came off by this morning. I know it's going to happen because I have already seen it happen.

Lost in all of my contrarian diatribe is the fact that I agree with a lot this. I like to weigh fairly often as well because it does help me understand what my body is doing. BTW, I also ate sushi on Wednesday and saw the resultant fluctuation yesterday. It didn't freak me out either because, like you, I expected to see that stuff.

I just don't think elimination of noise - on a day to day basis - matters much at all. When we are using a scale to measure, time is a much more important component (assuming we have our CICO in order).

I honestly don't know what point you're trying to make. It just seems like arguing for arguing's sake.

Silentpadna wrote: »

VintageFeline wrote: »

I honestly don't know what point you're trying to make. It just seems like arguing for arguing's sake.

It's not even meant to be an argument per se. As not just a numbers guy, but a math guy too, I'm not convinced that time of day matters as much as many think it does. That's the point. I see the recommendations to weigh yourself under the "same conditions", but don't see the logic in it for reasons I've tried to explain. I do understand the idea of doing that, but I believe the math is more complicated because of the number of variables involved. And because time is a monumentally bigger factor (all things equal), trying to make the conditions consistent or the same is not possible.

More consistent, less variables. Please don't make it black and white. It is better to improve the consistency as much as reasonably possible. If I can have a 5 lb fluctuation during the day, but can eliminate that from the data collection, that is preferable.

It's not perfect, it is better.

Silentpadna wrote: »

stanmann571 wrote: »

IF you weigh less often, you become more vulnerable to noise.

The more data points you gather, the smoother becomes the curve of your data.

Why is that? Doesn't it depend on the window of time you are taking data? What if you took 20 data points over a one week stretch, and I took 4 over an 8 week stretch. Which do you think would give you the most accurate trend? The one with more data points or the one with more time? Further, If I took 4 over an 8 week stretch and you took 40, would our trend lines be substantially different?

What is the smoothest curve? Either a linear function (no curve), or a perfectly functional curve where the the dependent variable (weight in this case) is precisely related to the independent variable (time). We know it's not, because weight itself is comprised of more than one thing and they change at different weights due to totally different circumstances. As a result, we end up with a trend line (or curve) that can be calculated using data points. If our goal is to lose fat we need a long time window to see the trend line because fat weight changes slowly.

Look at the data for stock prices or market indexes. Over the very long term the curve is much more smooth than shorter terms. Why? Because daily, weekly, or monthly fluctuations vary much more dramatically than long-term ones. If you are looking for what will happen to a stock price or index over a very long period of times, it doesn't matter at all what time of day you check its price on a given day. It's not a perfect mirror for weight fat loss, but it does illustrate that the more data points there are, the more noise there is. But...the number of data points don't affect the trend...they are simply components of it.

Here you go.

Representative data from my past months weigh ins. I've gained 2-3 lbs

However, had I only weighed 3 times I might believe that I'd lost 4 lbs.

1 237, 232, 235
2 234, 236, 235
3 233, 237, 234
4 234, 236, 236
5 233, 238, 235
6 234, 233, 237
...
15 235, 236, 237
...
31 233, 239, 237


And yes those are actual numbers from the past month. Not all of them.

In response to your stock market comment. the number that counts is the EOD number(which is actually the market close number) for the same reasons as weight first thing after voiding in the morning. It is the number least subject to manipulation by non-substantive factors.

I really can't figure out why you're pushing so hard since you claim to not even hold to the position you're defending.

Tacklewasher wrote: »

More consistent, less variables. Please don't make it black and white. It is better to improve the consistency as much as reasonably possible. If I can have a 5 lb fluctuation during the day, but can eliminate that from the data collection, that is preferable.

It's not perfect, it is better.

I understand the idea. But why is it better?

Whatever time you weigh yourself is cool. I'm not arguing that or debating that, and I'm especially not saying it's bad. What I am saying is that it doesn't matter much. Your own example shows a fluctuation of 2.5x (in a day) the amount of weight loss you would achieve in a week if you were aggressive, 10x if you are going slow. That means in order to minimize noise, you need time. Once you have enough time, the individual data points don't matter. Wherever you are in the 5 lb cycle of fluctuation throughout your day is irrelevant to your fat loss. Not only that, our digestive tracks move at different speeds for different people. Same with diet and workout type. All the different things that speed up or slow down fluid retention make it impossible to have a consistent baseline to work from in the first place.

Whether you eliminate it or not from the data collection is irrelevant. For example, I weigh 219 lbs right now. In 20 weeks I would like to weigh 189 or thereabouts, maybe less if I can, but I think I'll be slowing down my goal speed a little before that. Assuming I reach that goal, however many data points I get between now and then and when those data points occur is irrelevant.

Again, I am not saying weighing at a particular time is a bad idea. It's a fine idea. But because of the complexity of body processes, the magnitude of daily (and even hourly) fluctuations, the consistency seems to me to be a myth, even if it seems intuitively correct.