My So-Called Beeminder Life


I was at a weight loss plateau for the last couple of weeks of March. I thought about it, and realized I wasn’t cheating so much as I had gotten more used to the healthier foods I’ve been slowly switching to and eating more of them. For example I’ve discovered that I like avocados.

I noticed this and limited my snacking and I’m making progress again. The sledgehammer approach is to start using My Fitness Pal again and log everything I eat. It’s a fair amount of effort to do that though, as I have to enter ingredients for all the meals I prepare and I haven’t stuck with it in the past. Maybe I make a commitment to try that for a month if I get stuck again. Or maybe I just make a goal to be a little hungry a few times a day to cut down on snacking out of habit instead of need.


I was on a work trip from Monday to Wednesday; it was short enough and I had enough of a buffer that I didn’t bother setting up a break. I blasted through a weeks worth of cheat meals and that is being generous, but I also went to a grocery store and bought healthy snacks which I have not done before, so thinking about the goals definitely helped.

I’ve decided to archive my cheat meal goal and replace it with using My Fitness Pal for at least a month. It will take some time to enter data for the foods I eat now including measuring how much I’m eating, but I think my biggest issue now is that I’m still eating a bit too much and could make do with less. It seems worth it to pay close attention for a while and recalibrate.

I’ve set up a Do More Goal with the statement of logging everything I eat and coming in under the goal amount 5 days a week. It would be more, but I’m not sure what to do on days I eat out and just don’t know the nutritional data for the meals so I figure I will let those slide.

I’m not sure what to make of their idea that you should eat as many calories back as you burn exercising, especially since it feels like their numbers are on the high side for cardio but on the low side for strength training. I will just go with it for now, and remind myself that the real goal is just to become more aware of how much I am eating rather that hitting any specific calorie target.


So using MyFitnessPal worked better than I expected; my plateau ended immediately after I started to carefully log my food again. Some of it was just eating less because tracking everything forces you to think more, some of it was that my portion sizes had crept up. It is a nontrivial amount of work to enter in the recipes as I cook meals, but I now have them for most of my usual meals so it is faster. There is the problem again that if I cheat during a day then I keep cheating since I have already written the day off.

I’ve lost focus on tasks a bit. There were a few days where I wound up binge watching on Netflix and fell behind. Then once behind I’m likely to not even want to look at the overdue tasks in todoist so avoid it for another day or two. So I’m adding a new goal to not have any overdue tasks in todoist for 6 days a week. I looked at setting that up with IFTTT, but it looks like the only triggers are tasks done.

Also I’ve made a new goal for coming to work on time. We have flexible hours, but I’ve been pushing it a bit recently, and it would be strictly better not to. So I’ve put myself up for 4 days a week. This is one of those goals that would really like to not count weekends at all. Maybe I should make it 6 days a week and enter a point for weekends always?

To make room for the new goals I’m archive my “alive” goal to do more interesting things, and my “happy” goal to not get upset about little things. I’m far ahead on alive, and while my “happy” goal was useful, I think I have got most of the benefit I’m going to get out of it at this point.


I decided to upload it as-is 1. Let me know if you get any use out of it or if there’s any improvements to be made.

I also took the time to make my Anki Beeminder add-on fit for consumption 2! Hope this is useful to some people. It produces output as can be seen here 3.


Thanks, I’ll check it out. FYI, your anki goal is marked private, so we can’t see it. :slight_smile:


Cool! Now public.

I also updated the description of my Anki add-on. Will create a separate post about it in a couple of days, since a forum search revealed there to be a steady background interest in combining Anki+Beeminder!

If it wasn’t clear, my post about 3 months ago was to make myself accountable for sharing this deck and add-on. It worked after all :wink:


This plugin is fantastic, thank you!

I’m not sure if this is a bug report, or a self-inflicted error due to how I have set up my goal and/or Beeminder Sync settings, but:

I have Beeminder Sync set up to “use premium features” - “keep multiple datapoints (use with ‘last’, ‘max’, …)” - and my custom Beeminder goal is currently using the “max” aggday function.

The first 3 datapoints shown were added during the evening of May 20th local time, and the last datapoint was added at 12:30am (after midnight) on May 21st. Anki thinks of this time between midnight and my goal derailment time of 3am as the 21st (hence the different number of cards reviewed and duration), but Beeminder (or perhaps the way the Beeminder Sync plugin is sending the data?) still counts it as the 20th. Due to the combination of the max aggday function and multiple days’ datapoints being logged to the same day number, the 2.2 data point is effectively overwriting the 1.78 datapoint, and that 1.78 is now never going to get counted in the graph.

Am I trying to use this incorrectly? Should I go back to using aggday “sum” and manually fix up my duplicated datapoints, or change something else about my setup on Beeminder’s side, such as the derailment time or a different aggday?

(For what it’s worth, I often stay up past midnight due to my work schedule, and strongly prefer a 3am derail time instead of midnight)

What do you suggest? Thanks!


Deadline is support is the oldest part of my code, since that feature is absolutely essential to me[*]. :wink: That means I should probably revisit the relevant logic with some fresh eyes. What you are describing almost sounds like expected behaviour though!

I’ll check my logic tomorrow. Maybe I hard-coded my preferred deadline setting. Wouldn’t surprise me!

In the mean time, are you sure you haven’t set the “Next day starts at” setting in vanilla Anki? The way the add-on works, it should respect that setting and upload post-midnight datapoints as the previous day. I’d expect you studied 6 cards after midnight and thus the 1.78 is counted as part of the 2.2.

[*] fun fact, which I almost totally forgot: this add-on started because of a wager with the Beeminder devs, to create a solution that could upload my Anki time post midnight, as opposed to the Rescuetime goal I was using


Couldn’t resist taking a quick look at it before going out. You could try the hotfix version I uploaded! Hopefully I didn’t rush too much.


Aack! I didn’t even consider the case that I only studied 6 cards after midnight instead of 21…I think the bug is in my understanding about what was actually happening. Sorry about that! (I thought I studied an additional 21 cards for a total of 2.2 minutes after midnight, not 6 cards in 0.42 minutes, but I wasn’t paying very much attention.) I will try it again with more awareness about what is going on, but I have a feeling your explanation that everything is actually working as it should is probably correct.

FWIW, my Anki “Next day starts at” setting is “3 hours past midnight” (to have it match with Beeminder), but I believe the default is 4am. Given your asterisk comment, if you were using the default setting of 4am yourself and it was already working on the boundary cases for you (or at least within the midnight-to-deadline window), I’m not sure if that hotfix is actually needed…?


And to continue derailing this thread… I’m interested in this too. My feature request is for the ability to Beemind the number of new cards added. I see that everything else I want is planned already. Yay!


Hey! just wanted to say thanks a lot! this is very useful to me


I will definitely try out the Anki integration. I’m still playing with the French cards most days. It just doesn’t seem like the most important thing right now so it will wait for a bit.

It’s been a while: a bunch of things chained together. Some friends visited for a couple of weeks, and it turns out that having company including a 4 year old girl is highly disruptive to just about everything. I also forgot and didn’t set up breaks ahead of time, so burned through the buffer on a lot of goals.

Then after they left I had extra buffer from the belated breaks kicking in, and wasn’t feeling that motivated so I kind of burned through that for a bit, which then led into an on call week, which wasn’t actually that bad but of course I had set up breaks for that which gave more extra buffer to run through.

I fell into a pattern where I wanted to relax and throw myself into a computer game or Netflix or something, but the structure of goals to satisfy each day was nagging at the back of my mind preventing that. So I wound up fighting with the system and instead of a positive loop where keeping up made me feel more in control and inspired to do more there was negative feedback where I was looking at buffers as an excuse to skip things the next day. Last weekend I ignored all goals, spending Saturday diving into a computer game and Sunday doing some shopping not on my todo list and doing something creative did I go back to feeling inspired again.


I came back from a hiking vacation a bit over a week ago, having just gained a little weight that turned out to be mostly water weight and came right back off, so I am under 200 pound again. Yay! I’ve been going to the mountains with a friend from High School and his family going on 20 years now. They started as backpacking trips, but as of late they have strayed more into glamping territory. Now he has a 2 year old, so much more of the time is spent working around nap times and eating compared to actual exercise, but it was still a lot of fun.

Traditionally we always come of with some idea for a computer game we might want to make, although now I’ve pretty much resigned that there isn’t going to be much in the way of implementation while I still have a day job. Last year I came back with a different idea and make a Beeminder goal to spend 1 hour a week writing up the idea in more detail. It turned out to be a toxic task. I was inspired while on vacation, but when I got back into my daily routine I focused on other things and so would leave it. I was never in danger of derailing, but forcing myself to spend an hour being creative wasn’t enjoyable or productive, and it made me tend to avoid thinking about Beeminder and more likely to slip on other tasks until I finally archived it.

This year’s idea has a more achievable short term goal. With a little more work I can get to a paper prototype and try that with some friends. So I’ve made a new goal to spend 30 minutes a week on this. If keeping up ever becomes an issue I won’t hesitate to archive it.

Finally, I’ve decided to try a housecleaning service. I do just fine setting a 15 minute timer a day to keep up with small tasks, but if I do just that then things like cleaning the blinds which I realize have gotten quite dirty take forever. Being forced to clean up random clutter on the kitchen counters, and dining and living room tables before they come seems like a bonus.


Well setting a goal for 30 minutes a week on games didn’t work out. I fell off the wagon a bit again, and I’m not sure quite what went wrong. I’m essentially making another post now to force myself to think through things a bit more.

I am working for a different group at work since I got back from vacation late July through September. I knew that it would be more stressful, but it is also an opportunity to stretch myself and grow a bit, and I do like the change of pace. I think it’s a good thing overall, but being more stressed about work does cause other things to suffer.

I think having even around 9 goals is too many for me. I just stop taking them as seriously and it becomes a process of just entering data points. The novelty of Beeminder has worn off, so I don’t get quite the same feeling of inspiration from just entering data or looking at graphs, so when things are a bit stressful I sometimes disengage from it for a few days at a time.

The thing that is confusing about this to me is that I have internalized the idea that I need to be exercising and paying attention to what I eat. So even though I’ve gone from a week of buffer on my fitness and food tracking goals to almost derailing before my scheduled breaks set in, I’ve still lost a couple of pounds since my last post. Beeminder has clearly helped me get to this state, so it’s a success! So why am I fighting with it now?

I have a contrary streak, and self sabotage is part of it. I’ve caught myself thinking that since I’m doing so well in general I deserve a break here and there, which might be fine except that I know that I’m in a relatively brief period where I should be putting in extra effort so it’s a bad time to slack off.

Meanwhile my goals to keep up with tasks and keep my house clean I have not internalized. My sister and her family are visiting starting today and I managed to put off a lot of the final little prep things until the very last minute, which is exactly the kind of added stress I was supposed to mitigating by setting up these goals. Well, it turns out that setting a goal and then sometimes ignoring it and doing just enough to meet the strict minimum doesn’t actually help all that much.

Well now I have most things on a week break while family is visiting, and have set a few of them to be archived. (Feature request - I don’t want email reminders for goals that are on track to derail within my alert settings but after the archived date to bother me. I understand why that isn’t quite as simple as it sounds.) I will think about how best to reset after that I guess.


I was getting pretty close to cheating on a few goals, so I just reset and generally rearranged them. I guess technically that was cheating since for some I entered enough data to count for two weeks, then archived it before creating a similar item, but whatever.

I’m rather happy with the way the subscription plans worked out. I wasn’t a premium user, partly because you could get most of the benefit with manual fiddling with the road settings, and partly because the price point was a bit higher than I would prefer. My baseline is YNAB, which gives a nice app and has tons of community support for around $4 a month. I was happy to see the new layout, and many of the details of the subscription process are well thought out.

Also, somewhere along the line the details about road widths and the amount done has been streamlined. I remember being thrown off by effects due to the road width when I first tried Beeminder, and now when you start with 7 days of buffer in a goal you actually have 7 days, which is nice.

Anyway, I wound up getting rid of any timing in my goals, and replaced everything with the idea that I will set a timer and do some number of sessions a week. This also solves the problem of getting way ahead; I won’t enter more than 1 point per day on those.

Here is what I am down to currently.

Exercise, 5 times a week
My Fitness Pal logged and under calorie target, 4 times a week. (I think I will need to replace this with a more simple avoid junk food and overeating during lunches at work, but I will try leaving it at a doable 4 times a week now)
House cleaning: 4 15 minute sessions a week
Projects: 3 30 minute sessions a week on whatever long term projects I have
Todoist Tasks: 10 a week, some of these will overlap with above tasks
Overdue tasks: No overdue tasks in Todoist (It’s ok if I just schedule them in the future, but at least I have to look at them) 6 days a week.


For August and September at work I was on loan to another group with 2 junior-ish engineers that were understaffed for all the projects they wanted to get done by the end of September. This has been an interesting opportunity to act as more of a lead and jump into a bit of a different space and learn some new technologies.

At times it was pretty hectic and I wondered if I was going to be able to make all the pieces fit together, but overall things worked out pretty well, and people seem very happy that I helped out. It also had some of the aspects that caused me to get way too stressed out a couple of years ago that led me to commit to focusing more on my health and a balanced life. If there are some problems that I don’t feel are under control, I will spend a lot of mental energy looking at them from different angles all throughout the day. So even though I didn’t put in a lot of overtime it wound up really diverting my energies.

One of the reasons I agreed to it was that the deadlines weren’t really all that hard; we got all projects done but one, and that is launching this week while I am on vacation and no one is really upset about that slip. This was pretty key for me because I do terrible when I am dealing with an impossible problem; here I could always step back and realize that any problem that came up could not be all that bad. Also I had this week of vacation scheduled which was kind of a hard cutoff for the effort.

So what does any of this have to do with Beeminding? Well, there is good news and bad news. The really valuable part is that I kind of don’t realize how stressed out I am getting when I am in the midst of things. It didn’t help that the previous week I came down with a cold as well. Only when last weekend came and I had tied up all the loose ends I was going to tie up did I realize how tired I was. However I can clearly see that I had stopped cleaning my house, eating well, and mostly stopped going to the gym. Having those graphs is really very valuable; it makes it very clear the trade off I was making. I was offered a chance to stay in the group (with a more sane schedule, but still more chaotic than my previous position) and it would lead to a promotion quicker, so having the Beeminder data was useful as a counter balance.

Now for the bad news, and it is rather bad. Paying the monthly subscription has eliminated the line in the sand that was preventing me from cheating. The general rationalization is that I was too distracted to realize in time that I needed to archive or dial down the road. I decided that I really was going to prioritize other things and didn’t want to be distracted. It’s not about the money, a few of the goals were previously freebees that still were at 0, and nothing was at a high amount. If there was a one click button to just give up, get charged and have things derail and just stopped being bugged I would have taken that. The thing that I really didn’t want was to have my email inbox fill up with reminders that would tug at me.

So now I am in the awkward position of finding Beeminder valuable, but also apparently not believing in the whole commitment contact thing.

For starters I am experimenting with Complice as a change of pace. I have seen it highly recommended here a few times, and it seems like it’s worth it to try out that system. It certainly addresses my issue that I make goals, but then wind up doing some easier tasks and not making much progress on them. I think I will cut down on several of the Beeminder goals and rely on daily interaction with Complice for a bit and give that a try.

Of course on the 2nd day of Complice it said rumor has it that I was a Beeminder user (I guess since I mentioned Beeminder in how I found out about it) and walked me through setting up a Beeminder goal to set intentions every day. That was a very seamless integration. This goal is another example of my littler Beeminder paradox. It feels very useful for me to make this new goal. It’s a real artifact of a resolution I’ve made to myself, and it means that I can’t lightly decide to just blow it off. On the other hand, I can look forward to future me now, and if in a month I decide that Complice isn’t really the right tool for me, it won’t be a light decision and it will be one that I will want to take effect immediately…

Hmm, this became long. Sorry for the wall of text. And that I have fallen to the dark side.


FWIW (not that it’s your main point), I also agree that a “cry uncle” button would be useful psychologically, even if it’s not really any different functionally than the current implementation. For me, deciding “no, I’m not going to do that goal today, I’m going to just pay the price and get it out of my head” at a point other than 2am would make the choice to fail feel more deliberate, rather than “oops ran out of time, guess I suck”. (And I still haven’t found the magic frequency of emails that remind me to enter data, but don’t irritate me when I already know things are due)


I’d kind of forgotten about this thread. Since I’ve been using Beeminder in ways farther from how it was intended it seemed a bit less appropriate to chat about it.

The comment about emails is relevant to an experiment I started just this week. I wanted to get back into the swing of eating healthy after the holidays, so I’ve started a no sugar challenge for a week. This would seem to be a terrible use case, since of course I can just cancel it at anytime, but it is working for me. One problem is that the zeno reminding is unwanted.

I’ve set it up as a do more goal where every day I have to enter a point if I’ve avoided sugar, and I set it to have zero buffer. Well, technically I should only enter a data point before I go to bed, but in practice I do so after dinner because I don’t want to be overly nagged. I don’t see a way to have a daily reminder but not extra reminders on my emergency day.

If anyone has a better idea for how to set this up I’d like to hear it.

Oh, and I’ve set the slope to 4 days a week without sugar after the initial week. That will allow me to climb to 7 days of buffer over the next couple of months while still having the occasional cheat day. I’m a little worried that I will crave sugar again the day after having it. Maybe I will want to add a rule to say that if I fail 2 days in a row I start entering negative numbers.


If you set the deadline to be in the middle of the night rather than at bedtime, you may find a combination of deadline and ‘start nagging’ times that does what you want.