menstruation and weight fluctuations

This sounds a whole lot like anorexia to me… I’m kind of worried about this. I’d at least suggest consulting with some kind of medical provider or even eating disorder specialists to make sure that this isn’t going to cause problems.

4 Likes

Yeah…with respect @dreev, every time you talk about Beeminding weight and all these shenanigans you do to stay on track it makes my skin crawl and I want to reply-all with disordered eating resources. Dieting is such a fraught topic and the truth is dieting of any kind increases your risk of developing an eating disorder and/or disordered eating behaviors.

I’d also add, on the topic of menstruation, that a lot of people who menstruate experience cramps, headaches, and other things that are alleviated with ibuprofen and other NSAIDS. NSAIDs (especially ibuprofen) are most responsibly taken on a full stomach with lots of water.

I agree that Beeminding other things that are more directly in your control are better options. I like Beeminding active days and exercise.

I think some other psychologically healthy things to Beemind for weight loss could be eating mindfully, checking in on hunger before eating (is this hunger? boredom? unpleasant emotions?), or drinking enough water.

8 Likes

My intuition is that it’s entirely different. A hypocaloric eating disorder manifests as a kind of OCD where you obsessively undereat. (From what I understand; I have no experience with eating disorders, not counting the one also known as “being an American”.) I would expect Beeminder to be entirely superfluous for an anorexic. As for the risk of developing an eating disorder, it seems implausible to me personally but … just saw @southpaw’s response… yeah, time for me to shut up now since I don’t actually know anything about eating disorders!

Not at all. There are some interesting articles about QS apps and anorexia, both helping and hurting.

Some actual research/abstracts:

https://www.sciencedirect.com/science/article/pii/S1471015316303646

Here’s some anorexics discussing tracking apps:

https://www.myproana.com/index.php/topic/1720922-best-tracking-apps/

3 Likes

This thread is a great example of Goodhart’s Law.

Here are some great examples. My favorite: In Hanoi, under French colonial rule, a program paying people a bounty for each rat tail handed in was intended to exterminate rats. Instead, it led to the farming of rats.

This is also related to the QS principle - only measure meaningful numbers. Your weight is not a meaningful number at all. For instance:

One of the best things you can do for your overall health and to lose fat is to build muscle through progressive resistance strength training. I think everyone who is dieting, counting calories, or watching their weight should throw all that crap out the window and start lifting barbells instead. But a gain of muscle may actually increase your weight (in a good way).

As others pointed out, weight fluctuations can be due to things like how much water is in your body, menstrual cycles, male hormonal cycles, seasonal metabolic patterns, how constipated you are, your circadian rhythm, and who knows what else. These fluctuations are basically meaningless so it makes no sense to track them.

And then, of course, there’s the weight you gain from eating a ton of junk food and storing it as fat, which is generally not a good thing for your health.

So weight gain can be good, neutral, or bad, depending on the situation.

When you measure the wrong thing, and try to optimize it, you end up doing bizarre things like trying to dehydrate yourself so your weight sticks to an arbitrary Yellow Brick Road, or a repeating cycle of binging and severe fasting, and there’s no way this actually gets you what you want. You’re tailoring your behavior to a meaningless number.

3 Likes

Even if somehow (a) periodic short term starvation diets are a good way to control your weight and (b) somehow you are able to stay extra hydrated (important both that time of month and when losing weight) despite the fact that it makes your weight go up a lot on an hourly time frame and at the same time are most motivated to consume less food by avoiding the smaller amount that it makes your weight go up on an hourly time scale… Dealing with low energy and likely poor nutrition from fasting at the same as not feeling great to begin with and having special nutritional concerns doesn’t sound healthy. And don’t forget that you can’t slow down every month to deal with this like you can when you’re sick.

And (a) seems unlikely. The first Google answer result tells me that the average woman gains five pounds of water weight during her period. Losing more than one percent of body weight in a week is considered unhealthy. I doubt the average woman weighs 500 pounds. So losing five pounds to compensate for that water isn’t healthy. (Feel free to research more accurate numbers for this.)

(b) just seems a bit absurd.

2 Likes

To answer the original question rather than just arguing with dreeves: I gave up on beeminding weight. This is just the most extreme example that weight fluctuations on the timescale of days - and beeminder really only works on a timescale of a small number of days, that’s the point - mostly not being connected to actual weight loss.

The best way I can think of to use a beeminder weight goal is to use it in conjunction with an equivalent (and healthy) calorie counting goal, and if you have more than five pounds buffer in your weight goal you get to turn off the calorie tracking goal and try to eat about the same but don’t have to count.

1 Like

FWIW, I have a “weight” beeminder goal, but the value is a flat number set to higher than I’m ever likely to reach (although I did have to bump it up even more when I was pregnant). That, combined with changing the Y value to be within my usual ranges, gives me the benefits of data tracking and seeing my trend, without having the threat of beeminder encouraging me to take actions I don’t think would be the best for my long-term health.

I usually stay out of weight loss discussions, as I don’t have a horse in this race – my actual weight goals involve weighing more, not less. It’s pretty obvious to me that setting a beeminder goal to weigh more would be pretty trivially gamed without actually improving the metrics I care about (strength and health), and I feel it should be equally clear that beeminding losing weight is also pretty likely to not actually improve the metrics you care about (appearance and health). I agree with @zedmango that an entirely reactive weight loss system of fasting seems tailored to end up with you retaining fat and losing muscle, which is most likely the opposite of what you want. Your scale number is pretty close to meaningless (see posts like https://www.delish.com/kitchen-tools/advice/a53899/these-pictures-prove-weight-is-just-a-number/), and optimizing for it is going to get you equally meaningless results.

3 Likes

wow, thanks for all the passionate replies! I have created a para-goal for exercise and would like to create another one to eat out less (but am still unsure about subscribing). What I did for my last period was to set the goal to be much more forgiving temporarily… and email support to get a refund.

@dreev: disordered eating is complicated … glad to read your last sentence :slight_smile:

2 Likes

Thanks for all this input, everyone! Let me continue my bold new policy of not talking about things I don’t understand. Instead I’ll just argue that Beeminder in general is remarkably robust to Goodhart’s Law. And for me personally, getting on my weight loss road by dehydrating myself is what I call a loophole you can drive an ice cream truck through. See the “Wouldn’t You Just…” section of our Dirty Plate Club post. I dispatch weight beemergencies by fasting in a way that, from my understanding of the nutrition science, such as it is, promotes long-term health. Like I still drink water while fasting. And also by doing pushups and running up and down stairs and stuff, for the short-term goal of sweating my way onto my road but which is also making sure I’m not losing muscle instead of fat, since obviously I don’t want to do that.

1 Like

I believe your system works for you, but there are some necessary conditions that not everyone has

  1. A daily schedule that allows fasting and exercise first thing in the morning (That’s the only way I can see the sweating part work without also dehydrating yourself - hydrate just enough to safely exercise, get it over with, and then be hydrated for most of the day)

  2. No water weight fluctuations greater than healthy weight loss rates (this one is really lacking in the example in this thread)

There are probably others. Your solution is an interesting story about making beeminder work for something when it seems like it wouldn’t, but I doubt it’s good advice for many people besides you.

2 Likes

I’ve tried to avoid commenting in this thread, but I’m explicitly using Discourse to reply to bluetulip’s great response, so I hope it’s less gauche of me.

THIS, bluetulip.

My general heuristic is “Beeminder and a collection of tricks really, really works for me”, but weight is even moreso an individual thing. I really wish Beeminder worked as well for weight stuff for everyone as it does for me, but that does not appear to be the case.

Beeminder should be less stressful than not using Beeminder–that’s the whole point of it!

Additionally, folks, if Beeminder is causing you to be unsafe, talk to support, get off the goal, and figure out how you got in the pickle you got into.

3 Likes

A specific example of Goodhart’s law for a weight loss goal would be inducing vomiting just before the deadline.

I don’t agree with this part. (To be clear, I’m only talking about myself now.) If I have to wake up and leave the house like a normal person but am not on my weight road, it mainly means skipping breakfast and probably lunch as well. Though I can typically get away with a salad or something if social convention demands I eat something. And definitely I stay hydrated regardless. When I get home, if I’m not on my road yet, that’s when I may have to do stairs and pushups or something before I can eat dinner.

Also, if I know I’ll want to have a normal lunch the next day then I’m reasonably well induced to not stuff myself at dinner the night before in hopes of waking up already on my road for the day.

Yeah, agreed. I’m lucky in this regard. I do know menstruators who are also lucky in this regard. I just don’t know how to characterize what’s typical. I think I’d amend my original conjecture to predict that the kind of fluid retention that would thwart this is if you steadily retained more and more fluid for multiple days in a row. (I also know unlucky menstruators for whom this is exactly the case.) If the increase is, rather, over a single 24-hour period then, according to my very much untested theory, even if the retention then persists for days you should be able to counteract it via normal intermittent fasting without doing anything unhealthy. Assuming intermittent fasting generally is healthy for you, another big assumption that I don’t have the expertise to back up at all.

I’d say the fact that that never occurred to me proves my point about robustness to Goodhart’s Law! :slight_smile: Like @adamwolf says, if such thoughts are even crossing your mind, derail the goal and reply to the legit check to say you need out for health reasons.

PS: Couple other relevant forum links:

And this one, which I see I’m repeating myself from here:

And here’s me originally articulating a slightly more extreme version of all this 3 years ago:

I also just resurrected that thread to comment on a point about heavy vs light food, which seems like one of the stupidest parts of my whole concept. In fact, now I’m reminded of this public Facebook post by Eliezer Yudkowsky: "Proves Too Much" is the... - Eliezer Yudkowsky

I guess I’m making the claim that the Conservation of Mass Diet works as long as you adhere to it only via healthy actions. Which, yeah, I can see why this all sounds ridiculous to people. And yet it works for me. But I’m also appreciating that most of the debate is probably avoidable. I’m basically beeminding fasting and exercise as measured by my scale weight. You might as well beemind fasting and exercise directly, and just adjust the slopes of your yellow brick roads to keep your weight trending in the right direction, without having to worry about monthly fluid retention and such.

1 Like

I chose the bulimia example because it’s more clear-cut and more obviously disordered. But you specifically described yourself as going through repeating cycles of binging and severe fasting. That sounds a lot like binge eating disorder or anorexia.

Now, you also said that you’ve done research and are careful to fast in a healthy way. My concern, though (not specifically about you, but in general for everyone who might try this) is that mental illness is difficult to self-regulate because it involves the regulator mechanism itself failing. With anorexics or people who suffer from body image disorder, they genuinely think they’re overweight and they really believe they are just being healthy when they starve themselves or over-exercise.

I am not saying this is you, but I do think that anyone who beeminds weight and doesn’t eat until beeminder says so is increasing the chance of an eating disorder sneaking up on them.

Pretty much by definition, someone falling into an eating disorder is not going to be able to say “I better stop trying to regulate my weight for health reasons.”

Well, I think anyone trying intermittent fasting should be under a doctor’s care and see an eating disorder specialist at least once in a while. Messing with your body’s natural inclinations for when to eat seems dangerous.

I can understand beeminding intermittent fasting and exercise, but why bring your weight into it at all?

I’m not sure I understand when you say you’re basically beeminding fasting and exercise as measured by your scale weight - I don’t think this is accurate. Beeminding fasting and exercise would look like having a goal for “don’t eat after noon” and “do cardio 3x a week” and “lift weights 3x a week,” whereas your goals use your current weight, random fluctuations and all, to determine when to fast and when to exercise. That’s the part that concerns me…

2 Likes

Okay. It just occurred to me that my intuition on this is probably wrong because I tend to frontload my hydration. The order of my life makes it significantly harder to drink enough past mid-afternoon. So personally there’s no way I could drink my normal amount of water in the morning and lose weight during the day, but that doesn’t necessarily apply to everyone. Probably also depends on individual fluid retention fluctuation throughout the day.

1 Like

I also have no expertise, but I’m pretty sure it’s healthy for some (assuming we’re talking about less than a day) and not others.

It’s interesting to note that menstruating women don’t fast during Ramadan, and there are often practical reasons behind traditions.

1 Like

I think the ideal solution here would be to have the brick road automatically jump up on the certain days you trend high, if you can figure out the average length of your cycle, and then automatically bring it back down after the bloating/whatever wears off. A more intelligent trend sensitive yellow brick road. This would require some predictive analytics :slight_smile:

1 Like

For some, intermittent fasting is in alignment with their body’s natural inclination. For me, personally, I feel very good with a much-longer-than-average overnight fast (it’s a bit of a family trait, actually) and it’s trivial to extend that fast by a small percentage. (Personally, I can’t do the fasting until dinner that @dreev can, and I recall seeing some talk of research that suggests that, on average, women do better on slightly shorter fasts than men, but since I didn’t look at the actual studies there or get it from an expert, I’d want to dig into that more if I were curious. I just know that my body doesn’t feel good when I go past a certain number of hours.)

If anyone’s worried about whether something is healthy for them, I agree that talking to ones doctor is the right thing to do, but I would be reluctant to think of intermittent fasting as usually dangerous. Like someone’s already mentioned, there’s a fairly massive group of people who do it regularly during Ramadan. [Not that “lots of people do it” = safe, but just that we have a decent amount of information on best practices because of that.]

1 Like

All that ^ said, I still prefer to commit to something else (like the time I start fasting, or adding new healthy recipes to ones repertoire, or whatever) for myself, even though I’m fast-friendly. (Also, I wouldn’t personally want that to be my maintenance plan over the long haul, and I wouldn’t want to try to use completely different techniques for one and then the other.)

So, while I totally think that there’s nothing wrong with Beeminding weight in that way, I also think the choice of methods for nutrition intake or body composition change are extremely personal things because of the vast number of differences in the combinations of preferences, metabolisms, tolerance for length of fasts, blood sugar needs, schedules, etc. etc. etc.

To support all kinds of methods (including those who want to Beemind weight loosely, but not using the same method), I’d love to see something that allows even more of that individuality than the very large amount that Beeminder already supports. I wonder if there’s an aggregation setting that could be created for some people’s weight loss goals (and who knows what other types of goals this could be useful for) where, if the datapoint could simply be deleted and the road wouldn’t derail, it’s ignored and won’t cause a derailment.

That way you have more time to get back down onto your road, leaving some room for fluctuations, but there’s a certain day by which you have to be back on the road, fluctuations or not. And that would allow people who are simply not recording weight on days they don’t have to, to go ahead and record, for the sake of the data, if they wanted to. (I’ve thought of some additional details about this, but I’ll spare the people in the thread from an even longer post!) Since you can see your datapoints visually distributed, it reduces the chances of backing yourself into a corner, too.

1 Like