Love of data and math

Options
lmf1012
lmf1012 Posts: 402 Member
in Health and Weight Loss
For fun (yes, I am a bit odd!), I wanted to see how in line the math was for my weight loss over the last 5 weeks (I took out weeks 1 and 2 due to high water loss). Based on my TDEE and my actual calories in, my estimated loss for the 5 week period was 7.12 lbs and I lost 6.8. I thought it was rather interesting to see how close the math lined up.

I probably need some other hobbies lol

Replies

  • PAV8888
    PAV8888 Posts: 13,626 Member
    edited May 2021
    Options
    I think we need to add another category of predictor traits associated with successful weight management using MFP and similar:

    --subscribes to premium version 😹
    --posts on forums
    --enjoys playing with spreadsheets

    I'm two out of three! So far so good! *in conjunction with the intake and expenditure of appropriate amounts of calories 😇
  • penguinmama87
    penguinmama87 Posts: 1,158 Member
    Options
    I never got into numbers and data analysis until I was an underemployed college grad with substantial student loan debt. I thought I had a good handle on spending, until I actually built and started using a budget and suddenly I could track every single cent if I wanted. I got to see a lot of trends and patterns that before were only vague ideas in my mind, and I saw that there was a lot I could do to more efficiently pay off my debt while not feeling overburdened by it. I'm still using that same spreadsheet today, though now it's more detailed and I am doing super boring adult things like retirement contributions, but it helps my family live quite well. And I still find it fun!

    It only took me ten more years to figure out that I could apply basically the same principles to calories instead of dollars and cents. :D It takes harsh judgment out and it's just the facts. I find it incredibly empowering that the math checks out!
  • AnnPT77
    AnnPT77 Posts: 32,147 Member
    edited May 2021
    Options
    Yes, numbers and math are powerful in this context, as is (IMO) at least a very, very basic understanding of statistics. For this pursuit, health and fitness, it turns out that we *can* use algebra, and success odds improve if we do.

    Explaining the comment about statistics: I see so many people here not understanding that calculator/MFP calorie needs are just estimates, essentially an average, and that individuals will need calorie goals that range around the average: Clustered near average, for sure, but with some people surprisingly far from average, either high or low. (And that's if calorie tracking is reasonably accurate, let alone the math understanding that will help with that.) Same misunderstanding is embodied in a mistaken belief that a fitness tracker "measures" calorie burn rather than estimating it using statistical methods.

    Also, some folks are stressing hard about things that are arithmetically meaningless. (Example: 50 calorie difference between exercise estimates that come from different sources, when the person likely has a TDEE of 1500-2000 or more, and around a 500 calorie daily deficit - may be enough to care, but not enough to provoke major anxiety). Ditto for things that are statistically meaningless. (Example: One day eating 1000 calories over goal, in a month or more of compliance with a goal that's around a 500 calorie daily deficit, if not more. Realizing that it's just 2 days delay in reaching goal weight, roughly speaking, is calming, I think.)

    That's without getting into the ability to figure partial servings ("If 1/3 of my frozen pizza is X calories, how many calories is 2 slices if it's sliced into 8 pieces" . . . yeah, I know, weigh it, but y'all get my point. Even with weighing food, still need to know how to deal with a label serving size of 3 oz/85g when the portion you're eating is 115g or some such thing.)

    OP, I like thinking of weight management as kind of a fun, high-payoff science fair experiment for grown-ups, and for sure math is a part of that. I admit, I haven't done a detailed spreadsheet to narrow it down to an exact percent, but I'm one of the rare people for whom the so-called calorie calculators are not very accurate - I'm outlier-ish for unknown reasons, based on nearly 6 years of careful logging, including weighing food. Without some basic math understanding, that could've been a real obstacle . . . of a health-risk nature, in my case, since usually the estimates are 25%+ *lower* than my actual calorie needs.

    Now, with apologies for digressing, I'm going to pick on @PAV8888 a little (don't worry, we're friendly):
    PAV8888 wrote: »
    I think we need to add another category of predictor traits associated with successful weight management using MFP and similar:

    --subscribes to premium version 😹
    I didn't. (I have premium right now, didn't subscribe, long story.)

    On other threads, you seem to say that 5 years of maintaining is a reasonable criterion for weight management success, and I'm past that. 😆
    --posts on forums
    I for sure do, but I think your claim here shows observation bias: Of the many thousands of people who don't post on the forum, and aren't your MFP friends, how would you know whether posting on the forums is a success predictor, or a failure predictor via "get ticked off and leave"? Insufficient data! 😉
    --enjoys playing with spreadsheets
    Not a single one, unless you count trying out heybales' "Just TDEE Please" or consulting the fabulous protein efficiency spreadsheet over in the Foods stickies. Too "busman's holiday", for me. 😆

    I'm two out of three! So far so good! *in conjunction with the intake and expenditure of appropriate amounts of calories 😇

    Buying into that last one, big time . . . with a tip of the hat to OP, because IMO some math knowledge is good for an extra point. 😉
  • PAV8888
    PAV8888 Posts: 13,626 Member
    edited May 2021
    Options
    AnnPT77 wrote: »
    Now, with apologies for digressing, I'm going to pick on @PAV8888 a little
    You're so picky today Annie! These were additional correlations to consider! Not claiming causation or statistical significance, yet!
    AnnPT77 wrote: »
    PAV8888 wrote: »
    --enjoys playing with spreadsheets
    Not a single one, unless you count trying out heybales' "Just TDEE Please" or consulting the fabulous protein efficiency spreadsheet over in the Foods stickies.
    Well.... they're spreadsheets and you played with them, so your assertion of non participation is refuted! :wink:
    AnnPT77 wrote: »
    PAV8888 wrote: »
    *in conjunction with the intake and expenditure of appropriate amounts of calories 😇
    Buying into that last one, big time . . . with a tip of the hat to OP, because IMO some math knowledge is good for an extra point. 😉
    Well at least ONE item * discussed in my post was of statistical significance! :innocent:
  • middlehaitch
    middlehaitch Posts: 8,484 Member
    Options
    Oops, had to read the thread because people who are data nerds are exotic. B)

    Obviously not one myself.

    The only thing I do on @PAV8888’s list is post, and that right now is an occasional thing.

    Still highly successful at maintaining since 2009 though.

    Heck I just ran off to weigh myself and mid afternoon I was only 1lbs more than I was when I last weighed myself in February (usual first thing after peeing)

    I have a very relaxed outlook: don’t log, and a scrap of paper next to the scale with a handful of readings covering 18 months.

    (Eek, I could be swinging up and down 10lbs or so and never even know it. lol)

    For workouts I just have an average number for general, walking, and lifting, but can’t remember the last time I even used them. They work well if I start losing and have to log to adjust my intake.

    I have noticed there are lots of people here that do keep all kinds of spreadsheets, and I do admire them, it’s just not me.

    Cheers, h.
  • KNoceros
    KNoceros Posts: 324 Member
    edited May 2021
    Options
    I have an ever expanding spreadsheet with graphs and all sorts. I used to just use it to track Ci, Co, weight (and BMI) so I could compare “the maths” with reality and get an idea of whether my tracker was lying to me about co (it is - it reads about 150 LESS than my actual burn, so if I eat what it says I actually lose rather than maintain).
    It has since expanded to include steps, BP, separate sheets for logging and summing my workouts, and recording fastest run efforts. The graphs plot weight (with trend), BMI and BP over time.
    I have variants of this going back over about 6 years.
    I am sure it will grow and grow.
    🤓🤓
  • Antiopelle
    Antiopelle Posts: 1,184 Member
    Options
    I did a basic version for decades, but then when Covid hit, it took me some time to realize I was much more sedentary than before and my weight crept up slowly but surely. It was time to take data more seriously and I did a mini-graph next to my normal graph to really determine what my BMR and NEAT were and if I was tracking food and weight accurately. This helped to get me back on track. And it showed that tracker calories, kitchen scales and body scales are spoton.
    Love maths, love excel ! :smiley: This is what keeps me accountable.

    4js6iejebji8.png

  • viajera99
    viajera99 Posts: 252 Member
    Options
    janejellyroll wrote: »
    There was something so comforting to me when I realized that once I understood the math, my weight pretty much behaved exactly as you would expect for an average person!

    Exactly this! I have my own home-brew spreadsheet too with CI, CO, Net Calories, Weight, Predicted Weight, BF%, Fat Mass, Fat-free Mass, BMR & BMI. Seeing the numbers work out also lets me know that my calories-in are being measured with reasonable accuracy. It also keeps me from getting discouraged when the scale isn't moving, because I know it will catch-up eventually. In the Laws of Thermodynamics we trust.

  • Oishii
    Oishii Posts: 2,675 Member
    Options
    What I wish is that MyFitnessPal would do the stats legwork for me! It knows when I lose, how much I exercise (Fitbit) and how much I eat, so wouldn’t it be possible for it to work out my actual maintenance and loss calories on the basis of that?

    As it is, I have always lost far faster than MyFitnessPal predicts, no matters the method I use. I am currently 2.4kg lighter than I was a month ago, but from my diary I have eaten about 2000kcal over maintenance for that time period. I want the option to click a button and MyFitnessPal to do the maths of what my actual maintenance calories should be. Maybe it could even work out if I lose more than predicted when I exercise or if it’s my base calories that are out of whack.

    Obviously, the real error in my data might be my consistently overestimating my food intake, but that would be statistically so unusual from what studies say that I feel that is unlikely to be the whole answer.
  • sijomial
    sijomial Posts: 19,811 Member
    Options
    @Oishii

    It might be the "Box Hill Effect" - that's place just sucks the calories right out of you.
    :)
  • robertw486
    robertw486 Posts: 2,388 Member
    Options
    Over the years I would imagine quite a few people have made some type of spreadsheet or used other tools to look at all the data. Now with so many apps and so much data available it's going to happen. I used a couple simple ones over the years, but mostly for workout stuff over food in vs out.

    Having the data is a great thing for many.

    For those that do find some small error, it often brings to light the aspect that aren't exact. It can be exercise estimates, standard vs specific Atwater tables, overall diet, fiber counted vs not, water weights, etc, etc. If a person really wanted to they could make a mega spreadsheet, but for most of us the simplicity of a smaller one averages out most of the time.





    Maybe this is where I should seek out the best of the spreadsheet masters to help with another I'm working on. My spreadsheet foo is weak these days, and for some reason I've got one desire that has defeated me in a number of ways. And the worst thing is, I'm sure I've done it in the past, I just can't remember which of the methods worked!
  • ninerbuff
    ninerbuff Posts: 48,523 Member
    Options
    Everything in the Universe is math. Every element is made of atoms molecularly combined together. The count of the atoms will dictate the element.

    A.C.E. Certified Personal and Group Fitness Trainer
    IDEA Fitness member
    Kickboxing Certified Instructor
    Been in fitness for 30 years and have studied kinesiology and nutrition
    9285851.png
  • PAV8888
    PAV8888 Posts: 13,626 Member
    edited May 2021
    Options
    Oishii wrote: »
    What I wish is that MyFitnessPal would do the stats legwork for me! It knows when I lose, how much I exercise (Fitbit) and how much I eat, so wouldn’t it be possible for it to work out my actual maintenance and loss calories on the basis of that?

    As it is, I have always lost far faster than MyFitnessPal predicts, no matters the method I use. I am currently 2.4kg lighter than I was a month ago, but from my diary I have eaten about 2000kcal over maintenance for that time period. I want the option to click a button and MyFitnessPal to do the maths of what my actual maintenance calories should be. Maybe it could even work out if I lose more than predicted when I exercise or if it’s my base calories that are out of whack.

    Obviously, the real error in my data might be my consistently overestimating my food intake, but that would be statistically so unusual from what studies say that I feel that is unlikely to be the whole answer.

    MFP, knowing the activity multiplier you've pre-selected, still doesn't seem able to add net calories for exercise... so no, I don't think your dynamic adjustment is on the way quite yet! :wink:

    The nature of population estimates is that most people track close to them. Some track further away. A very few track quite a ways away. Both up and down.

    You seem to be one of the "unfortunate" ones who has to spend more on their grocery bill in order to feed themselves more than predicted in order to avoid wasting away :lol:

    You can always compensate by setting up to gain instead of to lose. Or you can give fake stats to your fitbit to make it predict you as needing more calories. Or you can tie your fitbit to a paint-can shaker when you sit down to compensate for any under-estimating it may be doing :wink:

    I've often considered tying the fitbit to the dog's tail and getting a good wag going by parading some neighborhood cats in front of her as an effective way to add some extra steps and calories when my fitbit was not cooperating with my hunger level!!! :wink:

    ETA: "no dogs, cats, or other animals were harmed during the filming of the above discussion"
  • AnnPT77
    AnnPT77 Posts: 32,147 Member
    Options
    Oishii wrote: »
    What I wish is that MyFitnessPal would do the stats legwork for me! It knows when I lose, how much I exercise (Fitbit) and how much I eat, so wouldn’t it be possible for it to work out my actual maintenance and loss calories on the basis of that?

    As it is, I have always lost far faster than MyFitnessPal predicts, no matters the method I use. I am currently 2.4kg lighter than I was a month ago, but from my diary I have eaten about 2000kcal over maintenance for that time period. I want the option to click a button and MyFitnessPal to do the maths of what my actual maintenance calories should be. Maybe it could even work out if I lose more than predicted when I exercise or if it’s my base calories that are out of whack.

    Obviously, the real error in my data might be my consistently overestimating my food intake, but that would be statistically so unusual from what studies say that I feel that is unlikely to be the whole answer.

    Theoretically possible, except for maybe the bolded, I think. (Bolded unlikely because reasons, but I'm too lazy to type a bunch of speculative stuff.)

    But this is the app that brings you a "in 5 weeks" prediction that's a head-scratcher, even assuming that all days would be alike, for the whole 5 weeks, which is a nonsense assumption.

    So, in practical terms, it's unlikely.

    IIRC, there was a guy (with the right background) who posted here about testing a web site that I think would do what you're describing, but for some reason his posts got a lot of negativity from the MFP user community. I thought it was interesting.

    But - speaking as someone for whom MFP estimates wildly incorrectly, in the same direction as you've found true for you - it's just not that hard to do the math myself, to get a reasonable estimate. (I got serious about that back in early 2016, when I was reaching goal; maintained since.) I haven't even spreadsheeted to pin it down more exactly, yet my back-of-envelope-ish personalized estimates were sufficient to let me predictably lose weight at around a pound a month, by intention, for many months, recently . (For me, MFP guesses something like 25-30% low, though I can't tell you *exactly* why. I think it's a multi-factor thing. 🤷‍♀️)

Categories