Feature idea: Link graphs for positive reinforcement

My overall impression is that Beeminder’s current incentive mechanism is primarily based on punishment. In other words, you get penalized if you do not do as you promised. There are minor exceptions to this such as the reward obtained from unblemished, perfect graphs, the social satisfaction of having visibly stuck to your goal etc., but these are all indirect. My impression, though, is that behaviourally, positive reinforcement might yield better results if used correctly. Those of you more knowledgeable on the topic, please feel free to correct me on this.

That being said, here is one way that I think could be used to inject positive reinforcement into Beeminder: Linking two graphs together so that safety buffer in one can be transferred onto the other. The best example for this is linking a do-more goal with a do-less goal, wherein if you do more than you promised on the do-more goal (most likely an undesirable task you have difficulty with), you get to indulge yourself more for the do-less goal (much more desirable, but something you want to limit yourself with). This could also work the other way around, so that if you manage to refrain from doing the do-less goal, you get to take a break from the less desirable task.

Now, such a link could potentially introduce quite a number of loopholes, allowing the user to play around with relative rates to effectively cheat the system. However, I think these could be addressed if a fixed ratio between the rates of the goals is chosen at the time of linking and enforced afterwards for road changes. For example, you could say “5 hours of productive work to earn 1 hour of tv”. Then, whenever you build up 5 hours of safety buffer on your do-more goal for productive work, you trade that for 1 hour of safety buffer for the do-less tv goal.

What do you think?

  • Uluc Saranli
7 Likes

Exciting ideas! I love thinking about the big picture of self-incentivizing like this. I’m hoping to see people set up things like this with the API to prove the concept before we think about this in Beeminder proper.

In the meantime, here are a couple links this reminded me of: @nepomuk’s Nine Greens and a recent comment on the SlateStarCodex subreddit:

For most of it’s users, Beeminder is a “stick”, but for a significant subset of users, it’s also a “carrot”.

For me, building up a safety buffer is basically “imaginary internet points”, only even better because they are actually useful (they give me permission to take a break from working on that goal), PLUS once Beeminder starts working for you, then the actual improvement to your life and the accomplishments Beeminder brings is quite a nice “carrot”.

That said, I get what you mean by “imaginary internet points” in the sense of Habitica (formerly HabitRPG). I don’t know if I want Beeminder to go full Habitica, but badges would be a nice middle ground.

PS: I guess what’s missing in the current API is any way to override the akrasia horizon, since fundamental to @saranli’s idea is having an immediate reward. So now I’m thinking that rather than work out the rules for safety buffer ratios for paired goals or whatever, maybe we could give out special akrasia horizon override keys and let people build an app that implemented things like this, with a promise not to abuse the power of the special key. Then with a proof of concept in the wild, we could build in the feature properly…