Boosting: a prediction market on ideas

The CredSperiment Getting Started
1

Overview

Boosting is a mechanism that allows people in the community to express that they believe a contribution is undervalued; that is to say, that a contribution is receiving too little Cred. In order to boost something, the booster must have Grain. They can then “plant” the Grain in order to Boost the contribution. That causes the Cred of the contribution to increase, in proportion to the amount of Grain that was planted.

The fact that Grain was planted is crucial; it means that Boosting is expensive, so that it serves as a credible signal that the contribution was undervalued.

The Booster has gone out of their way, and used their own Grain, to tell the community that something was undervalued. Therefore, they get a reward. The reward is a small share of the Cred that the boosted contribution earns.

For example, suppose you spot an extremely important Initiative which you believe will be very helpful to the project. Others haven’t yet seen its significance, so it doesn’t have much Cred attached to it; let’s say it currently has 10 Cred.

You decide to Boost it, planting (say) 10,000 Grain. Suppose this creates 100 new Cred for the initiative. Now that it’s worth a lot more Cred, someone is more likely to come along and Champion the initiative.

Becuase you played an important role in prioritizing the initiative, you now receive a share of its future Cred. If you were right that the initiative will be very impactful, you may earn a lot of Cred, and eventually, earn more than the 10,000 Grain you used to boost it in the first place.

Boosting creates a prediction market on the future value of contributions. It incentivizes people in the project to seek out under-appreciated contributions (or promising ideas) and then boost them to signal their importance to the community.


Boosting Use Cases

There are many reasons to Boost.

Speculative Boosting

Boosting promotes downstream engagement, but boosting also increases the cred that will flow to the booster. This creates a prediction market on ideas. Community members are incentivized to boost content that they feel is “valuable,” where “value” describes content they think people should (or will) explore. It’s a signalling mechanism for content curation with economic incentives built in. If a mechanism was created that rewarded early boosters/commenters more than later ones, it would incentivize more people to comb through the depths of the community discourse to surface the most interesting stuff.

Bounty Boosting

Boosting can also act as a bounty mechanism. This is possible via the initiative system. People can boost things that they want worked on, and then as people contribute to those things they will get lots of cred. This would be great for things like working on documentation or research/design related things. How these initiatives connect from Discourse threads to GitHub Issues is an open question, but it’s being developed. This will flow cred all the way from ideation to the shipment of an idea, rewarding everyone who contributed along the way (I specifically made a comment on this and a discussion ensued, but I can’t find it atm because there’s so… many… conversations… lol).

Threshold Boosting

This idea was just brought up in the Champions and Heros thread. Essentially, there’s a dichotomy between declaring championship and earning championship. Boosting allows someone to signal intent towards an initiative. From there, however, we want to incentivize ongoing and increasing participation. Threshold boosting would give a boost to contributors who pass the threshold contribution of an Initiative (say 20%). This incentivizes people to keep contributing as more people also contribute. This allows people to both declare that they intent to champion an initiative through boosting, but then also earn “champion rewards” for following through on that commitment and contributing throughout the lifetime of that initiative.

Futarchy Boosting

If you create an Initiative for a vote, and then need to stake Grain to vote in that Initiative, then you could have a system where the “winners” of the vote have their Grain converted to a Boost (staked into the Initiative) while the losers can exit. Since SourceCred is retroactive, this creates skin in the game where the people who stake the most Grain (and thus have the most voting weight) then have to tie their Grain to the outcome of a decision. If the decision results in future endeavors that have lots of Cred then that Cred will flow back to the decision and all those who boosted it with Grain. If the decision results in negative Cred, however, then Cred will not flow back to boosters and their Grain will essentially be burnt.

This would have a few benefits:

  • People are incentivized to engage in governance because if they contribute to discussions that lead to positive outcomes Cred will then flow back to the discussion around a decision as well as those who voted on the decision.
  • People who vote have skin in the game where their Grain is tied to the outcome of the decision. This makes decision making a prediction market on what will create the most Cred in the future.

Aligned incentives! Yay :slight_smile:


Boosting Resources

“Cred Historians” and Curators are the main agents who are expected to be using Grain to boost things.

SourceCred in 5 minutes explains Boosting from a high level.

SourceCred UI Design and Development explores how users might interact with the Boosting mechanism.

5 Likes
2

Boosting Mechanics

A few ideas on boosting have been discussed so far.

Speculative Boosting

  • Boosting promotes downstream engagement, but boosting also increases the cred that will flow to the booster. This creates a prediction market on ideas. Community members are incentivized to boost content that they feel is “valuable,” where “value” describes content they think people should (or will) explore. It’s a signalling mechanism for content curation with economic incentives built in. If a mechanism was created that rewarded early boosters/commenters more than later ones, it would incentivize more people to comb through the depths of the community discourse to surface the most interesting stuff.

Bounty Boosting

  • Boosting can also act as a bounty mechanism. This is possible via the initiative system. People can boost things that they want worked on, and then as people contribute to those things they will get lots of cred. This would be great for things like working on documentation or research/design related things. How these initiatives connect from Discourse threads to GitHub Issues is an open question, but it’s being developed. This will flow cred all the way from ideation to the shipment of an idea, rewarding everyone who contributed along the way (I specifically made a comment on this and a discussion ensued, but I can’t find it atm because there’s so… many… conversations… lol).

Threshold Boosting

  • This idea was just brought up in the Champions and Heros thread. Essentially, there’s a dichotomy between declaring championship and earning championship. Boosting allows someone to signal intent towards an initiative. From there, however, we want to incentivize ongoing and increasing participation. Threshold boosting would give a boost to contributors who pass the threshold contribution of an Initiative (say 20%). This incentivizes people to keep contributing as more people also contribute. This allows people to both declare that they intent to champion an initiative through boosting, but then also earn “champion rewards” for following through on that commitment and contributing throughout the lifetime of that initiative.
3 Likes
3

How to boost (UI UX)

This is an open question, esp since we would (presumably) want boosting to work on Discourse, GitHub, and potentially other platforms as well.

A few ideas…

Browser Plugin

Might be easier to maintain than platform specific solutoins. Would create a common UI and UX flow across platforms.

Discourse plugin

We could potentially write a Discourse plugin which keeps track of every user’s Grain, and gives them the option to Boost a topic from within the UI.

Bots

Could help us create a unified UX across platforms.

Example:

Related Thread

4

Fun Fact (rumor): Reddit upvotes were called “boosts” before they were called upvotes.

1 Like
5

All three types of boosting (speculative, bounty, threshold) I think are valuable and powerful. They should also be able to be used without interfering with each other (i.e. axiomatic).

Threshold boosting is fun because it feels like a game, adds that gamification element. Bounty boosting is a nice way to allow leaders a form of ‘top down’ way to direct of resources. Speculative boosting allows for a move open ended, robust mechanism. This is the one I’m most excited about. It may allow more “pure” game theoretic play. Here, price discovery and “trading” (via general interaction between “investors” (contributors aware of each others’ cred scores)), could be a powerful force.

Today I had a moment on Twitter today that made me wish speculative boosting already existed. See, I “discovered” an up-and-coming music video artist a couple months ago. An anonymous spaceman, Lil Bubble, first burst onto Crypto Twitter (CT) with a parody of Mad World (54k views),

I didn’t catch their “first show”, but did catch their second, bigger hit, a parody of Lil Nas’ All-Time Lows (155k views),

Fucking brilliant. My ‘buy signals’ were spiking. Catchy music, good iconography that tapped into “alt coin trading culture” (I labeled it in my head), cheap production costs.

Seeing their talent, and having a general desire to support artists in the space, I followed them, liked and retweeted all their subsequent videos (even ones I didn’t like as much), offered support in comments, even a little carefully phrased unsolicited constructive criticism. In short, I invested.

A couple days ago, Binance flew Lil Bubble to #BlockshowAsia2019, where they played the main stage and brought CZ up on stage. Presumably this was their first big payday (CZ wouldn’t fly artists to Singapore to play their conference without paying them).

Their followers have shot up to 11k in the last couple days. Future looks bright. And they’re probably taking their “series A” funding and reinvesting in their business/personal brand.

In a SourceCred world, Lib Bubble would have received funding earlier (they were a broke artist and hinted so until recently), and would perhaps already be making a steady living off monthly Grain payments. And I would own a small piece of their future cred. And I would likely be doubling down :heart: :rocket:

2 Likes
6

@s_ben, the use case you’re describing is actually where the idea for boosting started. I was thinking about music and remixes, supporters “buying” the grain in new songs and new artists, and then getting some flow of value as the songs/artists became popular and got remixed into more content. Even if the flow of value is just the ego boost of being able to prove that you were one of the first people to be into Kanye or something.

1 Like
7

Ah, cool. Makes sense. Copyright law is so restrictive and has such high transaction costs that something like this makes total sense. I bet if you got it up and running it would attract capital, but would not want to deal with the music business in any way. Heard too many horror stories.

8

Futarchy Boosting

If you create an Initiative for a vote, and then need to stake Grain to vote in that Initiative, then you could have a system where the “winners” of the vote have their Grain converted to a Boost (staked into the Initiative) while the losers can exit. Since SourceCred is retroactive, this creates skin in the game where the people who stake the most Grain (and thus have the most voting weight) then have to tie their Grain to the outcome of a decision. If the decision results in future endeavors that have lots of Cred then that Cred will flow back to the decision and all those who boosted it with Grain. If the decision results in negative Cred, however, then Cred will not flow back to boosters and their Grain will essentially be burnt.

This would have a few benefits:

  • People are incentivized to engage in governance because if they contribute to discussions that lead to positive outcomes Cred will then flow back to the discussion around a decision as well as those who voted on the decision.
  • People who vote have skin in the game where their Grain is tied to the outcome of the decision. This makes decision making a prediction market on what will create the most Cred in the future.

Aligned incentives! Yay :slight_smile:

+1 to @decentralion for brainstorming this idea with me today

2 Likes
9

Boosting Curves

In the current boosting design, if people boost something early they will get a percentage of future Cred that flows through that thing. This does not, however, reward early boosters more than late stage boosters. This means that Cred whales can sit back and wait for ideas to get traction, then boost them to get a percentage of future Cred. When they do, they will get Cred in proportion to how much they boosted. This misaligns incentives between “Cred hunters” who seek out promising undervalued/appreciated ideas and “Cred whales” who can then swoop in to take the lion’s share of the benefits once ideas are discovered.

For example: if someone boosts something early on when it’s still underappreciated then they’re getting a small piece of a small Cred flow. Then as the thing becomes popular a Cred whale will be able to boost it and get a larger share of a larger flow of rewards. The Cred hunter found the valuable contribution, put Grain at risk to signal to the community that the contribution was valuable, but then once the contribution gets traction and more Cred starts flowing through it a Cred whale would be able to boost the contribution 10X what the Cred hunter did and thus get 10X the Cred rewards. This favors Cred whales.

A boosting curve could reward early boosters with more rewards than later stage boosters. This would acknowledge the risk that early boosters take on relative to later stage boosters. This would encourage boosting early and often. The curve might look as follows where Y is a percentage of Cred flows and X is total amount boosted.

boosting-curve

To make this work in practice perhaps a certain percentage of everything could be reserved for boosting (20/80?) and then the boosting percentage (20%?) could be split among boosters based on the curve.

This is just an idea, but I think it’s a good one to align incentives between various levels of engagement and Grain holdings. Curious if anyone sees any problems with this (first conceptually because we’re in the design stage, then with implementation) and/or ways this can be improved.

1 Like
10

Looks like the Bitcoin curve:)

A curve could make sense, though to reason about this, I think we need to know the basic relationship between cred and Grain, which afaik we haven’t defined yet (please let me know if I missed something!).

When you boost something, do you own a “share”/% of the future cred? If so, then boosting early would give you a larger % of the lifetime. For instance, let’s assume 1 cred = 1 Grain. If I boost a contribution with 5 Grain when it’s new and only has 10 cred, I have a claim on 50% of all future cred. If a whale comes in when the contribution is at 200, they have to pay 100 Grain to get an equivalent 50%.

Or, alternately, let’s say you only have a claim on the percentage when you got in. Continuing the example above, when the contribution gets to 200 cred you have 5/200 = 2.5%. Here, perhaps you could have a situation you describe where whales scoop up the contributions like crill into their golden gills. But that also assumes that the original booster isn’t taking the cred they get from the contribution and putting it back into the contribution via boosting. Presumably if you keep “letting it ride”, you’ll build your “position”, causing an exponential increase in the amount of absolute cred you get from the contribution, similar to your curve?

If cred is released at certain intervals, is that new cred “up for grabs”? And at what interval is cred recalculated? Does this rate of release affect the compounding of gains?

This also brings up another question: what happens when someone acquires 100% of the cred for a contribution? Does this lock other contributors from boosting? Or is there a mini “market” for each contribution, where the more Grain boosted into it, the higher the “price”/valuation? Or does the graph remain the same, with the % cred you get from your contribution equal to the Grain you boosted / total Grain boosted from all participants?

I kind of like 1 cred = 1 Grain. Just for simplicity, but also compatible with a positive sum game…but also wonder if this is problematic or constrains us in some way I’m not imagining.

2 Likes
11

@decentralion wrote it down somewhere, but atm I think it’s based on the difference between lifetime Cred and weekly Cred. Will probably change in the future.

This is kind of the idea behind the boosting curve, but there would be a fixed percentage of Cred from each thing that would go to boosting. This boosting Cred would be split between all boosters based on how much they boosted and when they boosted. This means that 100% of the boosting Cred for something would flow to boosters at all times, and anyone who wants a piece of it can boost. The percentage of the boosting Cred that would flow to a booster would be determined by where the boosting curve was at the time of boosting. This means that anyone can always participate in boosting and when they do they will dilute the current boosters. Since boosting is a form of staking, however, the boosting curve would only only go in a single direction (unlike bonding curves that go both ways). This means that later stage boosters have a much higher cost of boosting than early adopters, inspiring “Cred hunters” to seek out and curate awesome stuff. Does that make sense?

While we would need a model to test out reinvesting boosting rewards to create a “compounding interest” type thing. In the boosting curve idea here, however, it works regardless of if something becomes popular overnight or over a year. Since everything’s relative and various communities will move faster or slower than others, a curve seems like a sure way to enforce aligned incentives. Really like the idea of reinvesting rewards to double down on boosting though. That’s really cool :slight_smile:

1 Like
12

Dumping this here as a reference point to be explored, discussed, and maybe incorporated into Boosting sometime in the future:

13

Ok, think I’m wrapping my head around this…Thx for the explanation.

As for the curve, I like the motivation behind it (incentivizing taking risks in new contributions (cred hunting) and decentralizing rewards/power). However, I’m always cautious when adding complexity. Especially for a system that’s already going to be difficult to explain. A linear curve makes intuitive sense. Also, one reason to avoid complexity is that it can often backfire and have unintended consequences. For instance, whales with lots of time, money and resources could be much better positioned to exploit complexity and outcompete plebs too busy working the fileds for their Grain to notice. Just like high frequency traders now essentially charge “rent” on the stock market by simply trading faster with specialized hardware and algorithms, one can imagine whales with sophisticated cred hunting algos that systematically buy up stakes in all fresh contributions before anyone else (perhaps within seconds), knowing they can “flip” a guaranteed percentage because some will “moon” (become important) over time.

I also think that if we make it easy (the default?) to reinvest harvested boosting cred in the contribution it came from, we might be able to get similar exponential effects. It would basically be an exponential curve with the rate being determined by the rate of cred being released, similar to compound interest. I know credit card companies have no problem compounding interest for me.

14

They would essentially be subsidizing the entire SourceCred community then, which doesn’t sound half bad. Also, to do so, they would need to buy Grain which would be priced on a curve too.

This would be great :slight_smile:

15

Good point :robot: :rocket:

There is something nice about the curve you propose, just because it’s the same shape as the Bitcoin curve. I just learned that that was created by Satoshi in a fairly arbitrary way: “He modeled the monetary supply as 4 gold mines being discovered per mibillenium (1024 years), with equal intervals between them, each one being depleted over the course of 140 years.” Nonetheless, it has proven itself as a core component (or so is believed) in Bitcoin’s proven model. Which makes it easier to explain to people, and easier for some to trust.

16

Always a win :slight_smile:

17

I also wanted to point out the option of “internal boosting” where only cred holders flow cred to the nodes, rather than paying to mint new cred at that node. This idea of flowing ones own cred to other nodes came up in discussions with @decentralion about flowing cred to community members who offered support, guidance, training, etc.

This could still be seen as an investment, where I divert my cred in the short term because I believe some initiative will yield far more cred down the line. We can use the same cryptoeconomic primitives for the grain based boosting, but do so for cred. This feels slightly more wholesome than opening up to speculators driving the direction of the project, no matter how wise they may or may not be. This also seems to naturally shrink the attack space to people within the community, which is already a huge win.

1 Like