Social Cryptocurrencies and Increasing Returns
To an extent, there is an innate human impulse to reward generosity with generosity, and vice versa. In communities, we can see this principle in effect in the gift economy. However, gift economies have trouble scaling beyond small communities. This is where digital social cryptocurrencies can be helpful. It would be interesting to design a currency which worked something like this:
- Someone does something helpful,
- This helpfulness is recorded in social currency,
- The accumulation of social currency makes receiving help more likely.
Such a currency could encourage generosity by providing the intrinsic satisfaction of recognition as well as making it more likely that others will help you too. It would encourage an abundance-based mindset in which everyone gains from giving.
The Basic Model
One way of achieving this is perhaps to design a social cryptocurrency such that, the more you have of it, the more it pays others social currency to help you. In other words, rewards are proportional to the reputation of the person you helped. We can imagine a simple model of how this might work:
To begin with, let’s assume a basic question-answer format, like the one found on StackOverflow. Users posts questions, and people try to answer. The original poster gets to tick which question they found the most helpful. Additionally, any member can up- or down- vote answers once each if they find them helpful.
Each of these social gestures has a payout in social currency for whoever receives them. However, we’ll define the payout self-referentially. For a tick on an answer by a user who asked the original question, we’ll award 1/10th of their social currency balance to the answerer. For upticks we’ll award 1/100th, and for a downtick, take away 1/100th.
Note: these are not transfers: new social cryptocurrency is created (or destroyed) in users’ balances. The person who performs the social gestures (ticking, voting) does not spend their social currency – their own balance stays unchanged.
What kind of effect would such a social cryptocurrency have on a community which used it?
An Example with Seven Dwarves
To get an idea, let’s imagine seven users: Bashful, Doc, Dopey, Grumpy, Happy, Sleepy, Sneezy. At the beginning, everyone has the same social currency – 100 units. Now Dopey goes online and asks a question:
Question: How do you make Apple Crumble?
– Dopey (100)
Next, Grumpy answers:
– Sleepy (100)
Dopey finds Sleepy’s link helpful, and gives him a “tick” to acknowledge it. This doesn’t cost Dopey anything. His social currency balances stays at 100. However, now that Sleepy has received a tick, he earns a unit of social currency. According to the definition above, that’s 1/10th of Dopey’s social currency balance earned:
100 / 10 = +10
Sleepy’s new balance is 110. Now, suppose he now asks a question:
Question: What’s the cure for insomnia?
– Sleepy (110)
Obviously, because of his last answer Sleepy has a higher score. If another user manages to answer his question satisfactorily, he could earn slightly more social currency than usual. Perhaps as a result, the question begets multiple answers:
Answer: There is no cure for insomnia, fool. – Grumpy (100)
Answer: Try getting more exercise, and eating less junk food. Also take up meditation and do breathing exercises before you go to sleep. Don’t stay up late at night browsing the web. – Doc (100)
Answer: Doc’s advice is great. Perhaps you could also try a New Age meditation tape. I find them most relaxing myself. – Happy (100)
As a result of his unhelpful slight, Grumpy gets voted down twice, by Sleepy and Happy. The total loss to his reputation is 2.1:
110/100 = 1.1 (From Sleepy)
100/100 = 1 (From Happy)
His new balance is 97.9. Let’s assume Doc gets the tick of approval. His balance now increases by 1/10th of Grumpy’s score, or 11 points. He is now on 111. Happy’s tip earns up an up-vote by Doc, worth 1. His new balance is 101. The outcome of this round of the game is:
- Doc: (111)
- Sleepy (110)
- Happy: (101)
- Bashful: (100)
- Sneezy: (100)
- Dopey: (100)
- Grumpy: (97.9)
Doc leads the reputation charts because he provided a helpful reply to Sleepy, already well-regarded for pointing Dopey to his Apple Crumble recipe. Obviously, as this game continues, we can expect to see the helpful people ride to the top, and the less helpful to fall back down to the bottom. Those who are the most helpful are more likely to be helped, because doing so pays more reputation. Them that has, gets.
Some (very general) predictions
What would likely result from this game going through another 100 or 1000 rounds? To answer that in a detailed way, we might need to actually simulate it on a computer. User-agent modelling can be one approach for learning about the dynamics of such games, and their likely outcomes given different starting parameters.
However, without going to that level of sophistication, I think there are some basic things which are likely to be true:
- Users who are the most helpful will lead in terms of social currency, while the least helpful will follow.
- Those users with high scores will attract more help from others.
- The rankings will move more frequently the larger the payout for a tick/vote as a proportion of overall score.
- Inflation will be dealt with, at least in part, by the fact that the value of one’s social currency is always relative to other people’s, not absolute.
- All things remaining equal, it never gets easier or harder for a newcomer to earn trust, since the amount of social currency they can earn is always proportional to existing balances.
- Average balances are always increasing.
The last point is a bit problematic. If balances are always increasing, social currency scores will start to increase exponentially, quickly becoming unintuitive to the community which creates them. Perhaps this could be remedied by introducing a decay function, which destroys social currency due to inactivity over a period of time. In that way, maintaining a stable level of social currency would require constant effort. This is not a bad idea, since it reflects reputation dynamics in the real world, anyway.
Hopefully, this post gives a sense of how social currencies might evolve to encourage positive feedback, and increasing returns. I haven’t gone into too much detail yet – perhaps the detail will reveal the idea to be unworkable. This model deliberately leaves out a lot of potential nuances in the way social currency is awarded, and doesn’t consider alternative models beyond question-answer format. However, with some imagination the idea could perhaps be extended in these ways too.