Great Autoratchet Flippening

Bugabee
1

I saw the “Great Autoratchet Flippening” mentioned in the last daily Beeminder mail. I have not heard of that before. What is it?

1 Like
2

I found it see below:

2 Likes
3

thank you for finding and sharing this, @linux535! (i’m only on the weekly beemails, daily feels a bit much…)

i’m absolutely shocked to learn about this XD i get it, it makes sense business-wise and philosophy-wise, but personally, i will need A Lot of warning before this gets implemented! (mostly because of my uni goal, which currently has 34 well deserved days of buffer. capping those at 7 would feel absolutely brutal! all my others are usually below 7 / i wouldn’t mind them automatically capping at 7.)

4

Automatically trimming the buffer to seven days would mean I could no longer use Beeminder for minding my weight - my weight goal has a very gentle downward slope, so it quite often has months of buffer. Then I meet up with friends and suddenly the buffer is gone.

I have several other goals that I would delete if the buffer was auto-capped at one week.

I think making the current pay-to-use auto-trim feature a free feature is a great idea, but automatically capping buffer would be very impactful (for me at least).

One month would probably not be a problem for me, but I don’t really see the point of implementing that.

1 Like
5

Yeah, one of the first conclusions we came to from floating these ideas in the daily beemails was that a 7-day autoratchet is too restrictive. 30 days is feeling reasonable to most people.

Let me actually just quote the latest daily beemail to serve as an update on where we’re at with autoratchet flippening:

I’m getting lots of passionate replies about the Great Autoratchet Flippening. It’s great!

The first conclusion is that 30 days is better than 7. Lots of you have lots of goals that happen less frequently than once a week. Less frequently than once a month seems rare, though such goals exist. One user has a goal for deep cleaning their car a few times a year, for example.

Another great point from one of you: To calendial properly you need to be able to accumulate at least 14 days of buffer.

So that’s one question settled.

I’d also ideally like explicitly scheduled breaks to always be immune. Of course that’s opening a can of worms that could really slow down getting this shipped. But the idea is that if your datapoints are soaring far above the red line, you’re not really beeminding. You’re just making a graph. Not that there’s anything wrong with that, but that’s in the category of features that it still makes sense to me to have paywalled, despite our pledge focus.

Scheduling a month-long break though, that’s completely in the spirit of beeminding. Or even a break for the whole summer. Anything you’re doing very intentionally is great. Picking a goal like 50 steps per day from your Fitbit and ending up with years of safety buffer… that sadly happens all the time, isn’t intentional, and really defeats the point of Beeminder.

One more question for you all. Suppose that, for consistency, every goal had to have a finite autoratchet setting. Grandfathered goals or premium users could bump that cap to a maximum of 365 days. New goals on the free plan could set it to at most 30 days.

STRAW POLL: How would you feel about that?

Replies so far indicate some unease about possible cases (like a yearly flu shot) where you may want (slightly?) more than 365 days of safety buffer.

There are also questions about scheduled breaks and zero-rate goals like staying at Inbox Zero forever. Stay tuned! And keep the questions and feedback coming!

3 Likes
6

At least in that case, an extra break can be added closer to the deadline in order to push it out, even to precisely target it becoming due on the day you’re planning to get the shot, since there’s acres of time to play with… :thinking:

3 Likes
7

an extra break can be added closer to the deadline in order to push it out

My first thought in response to that is it defeats the purpose of “set it and forget it” which is no small part of why I love Beeminder. I don’t have to remember when to do something; I just let my goal remind me. Adding the extra overhead of setting a break in order for a once-a-year goal to work would be absolutely terrible for me personally.

Another thing to consider on this auto-ratchet topic is weight goals where you’re not working to gain/lose. I have over 76 years of buffer on my weight goal because all I’m doing is tracking the data. This objection would also apply to people who are maintaining or to people (like a previous commenter) who very deliberately keep a very slow decline.

3 Likes
8

Here’s my reply:

I don’t know why someone would have an autoratchet of 356 days. I don’t know why you want to have a limit to the autoratchet on premium accounts. You might end up making a choice that limits one specific user from attaining a legit use case. But I don’t know. I don’t feel super good about this. I pay for the bee, I want the whole bee!

Beedialing is hard to get right IMO, there is still stuff that falls through the cracks.

It’s a useless complication to a system that already works well. I only see bad arguments for the 365 days switch.

Yes, you’ll have to calendial every year… But! having a 365 days max autoratchet will give you X. X is unknown

2 Likes
9

For this “once a year” flu shot reminder use-case, it’s actually unclear to me how you would handle this by just “set and forget” anyway, so perhaps you could describe the situation a little more? Like, for a flu shot specifically, I don’t think it’s very likely that I could just go and get the flu shot on the exact day my goal becomes due, without warning. Appointments are usually required, or time off work, or other logistical concerns. It’s not something you can just set and forget, and then execute on the spur of the moment.

Here in the UK, my method would be to set the goal up to 1/year, and keep 365 days of buffer. On the day I book my flu vaccine, +1 on the goal. 365 days later, it will come up again for me to book my next appointment. Once done, +1 on the goal, safe until next year.

If I was planning to do it via drop-in vaccination centres and assume that I can just leave work and get a lift and all the other logistical stuff that would have to fall in line, so I want to beemind the actual act of getting vaccinated, then I’d probably use lead days to give me a reminder X days before the goal is due. I’d book the shot in response to that prompt, adjust my goal with a break so it would become properly due on the right day, and then add +1 on the day I get the shot.

You could also do it via fractional beeminding, where booking the flu shot gives you 0.1, or whatever. Next time the goal becomes due, the appointment isn’t yet, so you add 0.1 to acknowledge you’re still on the ball about this. Next time the goal becomes due, it’s your appointment date and you add the 0.8.

This is also a concern of mine. Here’s roughly what I sent Danny on the subject (I think this may actually be my first draft and can’t be bothered to check now, but the point stands):

“[Autoratchet] forcing me to fail on my weight goal just because my medication messes with my weight would not be helpful; I cannot control some of my intense weight fluctuations, short of not taking my medication (which is obviously a stupid idea). I can end up losing weight quickly but unsustainably, and then the goal would ratchet and the (inevitable) upswing would derail me. As it is, I’m being constrained to a gentle downward trend with lots of room for those annoying-as-heck fluctuations, and this is how it needs to be. I’ve seen it eat 90+ days’ worth of buffer in one horrible day of medication effects (and take weeks to go back down).”

1 Like
10

I’m the one who originally mentioned the thing about a goal to get a flu shot once a year. I do have such a goal, and I look at my goals often enough that I see when it is coming up within the next few weeks and it reminds me to sign up for a flu shot clinic at my workplace or schedule an appointment at CVS or whatever. Then when I get the flu shot it might still be a week or two before the deadline, and at that point I want more than 365 days of buffer since I don’t want the deadline to move earlier next year.

Honestly, I’m not that concerned about this; as others have pointed out there are better ways to organize this (such as having a goal to make an appointment or whatever) and it’s probably not a very common use case. (Also I have a premium account so it wouldn’t affect me personally in any case.)

2 Likes
11

I don’t have particular feelings about autoratchet being premium/free, but since @dreev mentioned the “spirit of beeminding”, here is what I said to him in an email:

Ok, so a question for you. I have a “walking” goal (to walk/hike at least 4 km per day) which right now has a buffer of 80 days. I used to have a buffer of 90+ days, but then covid happened and it dropped to 30+ days.

I know I could dial the goal down during the pandemic/illness/whatever, but I have an explicit reason not to. I really want to maintain the average (over many years) to be 4+ km/day, and setting breaks/dialing down would at best make it difficult. I am also metaminding (kind of) this goal – my “productivity score” goal (where I assign myself points for doing various stuff I have planned) has an entry for “achievements” (it’s reset every day), and one of the achievements is “walk 4/6 km this day” (it has two “levels”).

Is that in the spirit of beeminding?

Another example. I have a do-less goal called “sweet days”, where I count days I allow myself to eat sweets. It has a rate of 0.1 (so one “sweet day” every 10 days) and ideally I want to have a buffer of about 100 days (so that I could theoretically eat sweets 10 days in a row – not that I want to!). But even without such craziness it’s very good to have a buffer for this, because (a) sometimes I may be invited somewhere etc. and be offered something tasty, (b) some “sweet days” are planned (birthdays), and there may be two of them in a short succession, and (c) sometimes I want 2-3 “sweet days” in a row (e.g., Christmas). Add all that together and a buffer of 3 “sweet days” (which means 30 days) is suddenly very small.

3 Likes
12

i’m normally a big fan of consistency, but i don’t think it’s needed here! an arbitrary 365-day-autoratchet on every goal seems to only complicate things for a few select goals, as pointed out here. i don’t think it’s worth it. i assume the people setting 50-steps-fitbit-goals and having decades of buffer are mostly on the free plan, anyway. i personally would probably prefer it like this:

  • subscribers get no extra autoratchet, because they’re already using it anyway (or at least have that option).
  • new free goals get 30 days autoratchet.
  • grandfathered free goals get the option to opt-into 30-day-autoratchet. (so that current weight goals etc. wouldn’t be affected, but i could autoratchet all mine if i wanted to.)

also i think i just fomo’d myself into going from weekly to daily beemails XD let’s see if i end up regretting it and dial them down again eventually XD

1 Like
13

One more question for you all. Suppose that, for consistency, every goal had to have a finite autoratchet setting. Grandfathered goals or premium users could bump that cap to a maximum of 365 days. New goals on the free plan could set it to at most 30 days.

This seems the best to me. I actually wanted auto ratcheting but was worried if this change would remove flexibility of choosing what I wanted the auto ratchet to be. I’m in favor of things that encoruage me to stretch myself and get more out of beeminder (though I still don’t want to pay…)

I do have several goals which are ‘graphs only’ but agree that making that a premium feature is fair. (I have them for weight as a complement to my eating/exercise goals since I want to track inputs not outputs but still want the graphs.)

1 Like