Lying, cheating bathroom scales

I came across this from a blogger I like:

Andrew Gelman reweighs himself on his bathroom scale 46 times to compute the standard error. I mention this partly because it is inherently cool, and partly to tell the story that you cannot do this on my bathroom scale. If you do, you will get an answer of zero. It will come back the same every time.

Is that because the scale is super accurate, or at worst off by a fixed amount? Oh, no. Nothing like that.

It is because someone decided that the scale should have memory. If it gives you 161.3, then it has decided that everything from about 160.9 to 161.7 is going to count as 161.3 for a while. You can even see it, sometimes, bouncing towards the ‘real’ number, then at the last moment it reverts to its baseline. So if you (for example) were to pick up something weighing 0.2 pounds before weighting yourself, then weigh yourself again without it, you’d get an answer 0.2 pounds higher than otherwise.

I am fascinated by who thought this was a desired behavior. Writing this inspired me to get a second scale, for now keeping both around because it is fascinating.

I call that an unforgivable anti-magic violation and want to know what scales are doing that. I’m pretty sure Withings has never done that. (In fact, I rely on it not doing that so I can tare myself [1] with and without clothes in order to track my weight throughout the day.)

Can anyone confirm or deny for other scales?

Update based on replies:

Good Scales

  1. Withings (vouched for by dreev)
  2. Renpho (vouched for by Zvi)
  3. Greater Goods (vouched for by Don’t Worry About the Vase commenter)

Bad Scales

  1. Vitafit
  2. Etekcity
  3. EKS

[1] Example:

5 Likes

I know this has happened in the last decade in consumer weight scales. I spent a minute and tried to find more details, but this is the best I got:

We learned that some scales offer a seemingly manufactured sense of precision and accuracy: They may calculate your weight based on the memory of your previous readings rather than actually weighing you each time.

For example, one tester—after weighing herself, then weighing herself while holding a 10-pound dumbbell, then weighing herself yet again without the dumbbell—found that one of the scales said she weighed 10 pounds more than she did. Other scales would at first correctly show the additional weight difference when another tester was holding one book, but would then revert to her previous weight measurement without the book, suggesting that the scale doesn’t pick up on subtle weight fluctuations.

https://www.nytimes.com/wirecutter/reviews/the-best-bathroom-scales/

4 Likes

The scale I had before I got a withings scale definitely did this. If I wanted to get my exact weight after weighing myself recently, I had to trick it by partially standing on it to give it a weight far enough for my previous reading to reset it. Unfortunately I don’t have that scale anymore so I can’t give you the brand or model.

EDIT: Never mind! It was in my Amazon order history. Here’s the culprit:

Etekcity EB9380H Digital Body Weight Scale

amazon.com/gp/product/B00F3J9G1W

3 Likes

My scales do this as well. I’ve been meaning to replace them with something that can measure more reliably for way longer than a decade. EKS branded, I found them in the 1999/2000 Argos catalogue which is about when I bought them I think.

3 Likes

This is fascinating indeed.

My kilograms are not Duolingo points, so it’s a good feature, right? I should not check my weigh every five minutes anyway. Feedback loop should take some time.

In other words, there’s no such activity as “weighing items”, there’s just “losing my weight”. What is the purpose on checking on atoms count? There’s only my well-being relative to me from yesterday.

2 Likes

Can you imagine if your weight was like Duolingo points? I’d shudder if I found out that green demon was involved in my health–and I have a 3000+ day streak! :slight_smile:

1 Like

Mine does that. Vitafit VT1703U.

My last scale used to give a bad first measurement first thing after taking it out of the closet, so I bought this one because of the reviews saying it was accurate. Well, in my tests, it gives the same readings as the last one, but has this annoying habit of never disagreeing with a past measurement, so I’ll never be able to tell if it ever starts to break in the same way as the last one.

3 Likes

Aha, infuriating. In case anyone is just skimming, the conclusion so far is NEVER BUY a Vitafit scale nor an Etekcity scale. UPDATE: I’ve added a good/bad list to the top-level post here for reference. More examples solicited!

2 Likes

Ooh, Zvi has linked back here and he and commenters have new scales to add to the list. I’m editing them into the top-level post here.

Zvi’s commentary: “A vision of one form of nightmare awaiting us in a future full of things that have ‘intelligence.’”

2 Likes

I have the same scale and readings suddenly make so much more sense now that I know it’s doing this. I’m particularly outraged because it was (formerly) a wirecutter rec.

3 Likes