Future Salaries: Predict Prediction Markets or How to Hire Google to Account Salaries for Scientists and Free Software Developers

The App – register now to receive a salary!

See also:

Prediction Markets can be defined as markets for future events. Such platforms make it possible to obtain, aggregate and process information of dispersed knowledge. Two different contract categories require differentiation: binary (reveal probabilities of occurrence) and indexed (reveal mean values of the underlying index). Both payoff types are present in the analysed dataset of this study. Binary contracts are stocks tied to events that either occur or do not occur at a specific date or time interval. Contracts pay off $1 (or $10, $100, respectively) in case an event happens, zero otherwise. A hypothetical example for a contract is whether or not the United Nations will impose additional sanctions against Iran before 31.12.2010, paying off $1 if the event occurs according to predefined sources within this interval; zero if it does not. A market price of $0.67 suggests that the last trading occurrence between traders implies a 67% probability of occurrence.

Indexed contracts pay out the corresponding value of an underlying indicator (e.g. stock prices, interest rates, exchange rates, GDP, industry-specific data, etc.) at a specific date or the mean over a specific range, respectively. A contract paying off $1 for each rounded thousand-digit of the Dow Jones Industrial Index at 31.12.2010 is an example for an indexedcontract/event. Assuming a value of 12,230 at expiry date this contract would pay off $12. The ex-ante market value (e.g. $8.71) reveals the market’s mean value of what the market believes that value will be upon expiry. [2]

In this article I will consider exclusively “quantitative” (rather than yes/no) (for indexed contracts above) prediction markets.

“Conditional Tokens are a new, application-agnostic, asset class designed to facilitate the creation of highly liquid prediction markets. They enable combinatorial outcomes for high resolution information discovery through prediction markets.” [3] That is conditional tokens is the blockchain way of facilitating prediction markets.

So condition tokens are guaranteed by the cryptographic logic of the blockchain to be exchanged for some other asset (collateral) proportional to the index values at the time (to be precise, after the time) of the event.

In fact, there can be multiple collateral currencies, but that does not matter much for our discussion, as in a free stable market they are just different representations of the same measure of value.

In this article I will consider exclusively “quantitative” (rather than yes/no) (for indexed contracts above) prediction markets.

“Conditional Tokens are a new, application-agnostic, asset class designed to facilitate the creation of highly liquid prediction markets. They enable combinatorial outcomes for high resolution information discovery through prediction markets.” [3] That is conditional tokens is the blockchain way of facilitating prediction markets.

So condition tokens are guaranteed by the cryptographic logic of the blockchain to be exchanged for some other asset (collateral) proportional to the index values at the time (to be precise, after the time) of the event.

In fact, there can be multiple collateral currencies, but that does not matter much for our discussion, as in a free stable market they are just different representations of the same measure of value.

Particular examples of social good prediction markets

  • the number of downloads of a given open source software package (need an anti-Sybil protection to avoid manipulations)
  • the number of software packages referring to a given open source package
  • the above mentioned number adding (probably somehow weighted, it is a topic for further discussion of how to weight it) the number of other packages referring to them directly or indirectly (in the simplest variant the number of all direct and indirect references)
  • the number of citations of a given scientific (or other) article
  • and (likewise to software packages) counting indirect citations

Self references/citations probably should be excluded (but counted in indirect paths of others’ references/citations) to avoid manipulations.

This would allow to pay to a software author or a scientist accordingly to the predicted number of his future references/citations.

It turns open source and science into a business. The line between a business and a charity is blurred this way, so in the future legal system (if any) should be probably no distinction between a business and a charity.

Prediction of prediction markets

The main trouble implementing this idea is the following:

There are no prediction oracles for many important potential prediction markets:

  • no reliable public data about the numbers of citations
  • no aggregate data about software package interrelations

The best way to solve this problem seems to stimulate the market to create such prediction oracles in the future by providing money to the creators of oracles. (Hey Google, we are hiring you!) This can be done by prediction markets themselves: To kick up the system we initially need to create just one kid of prediction markets, the market of values of prediction markets.

We would have a prediction market for any kind of prediction markets, for example:

  • a prediction market for the value of software packages interdepencencies prediction markets
  • a prediction market for the value of scientific articles citations prediction markets

Furthermore we would have different prediction markets for different kinds of counting (e.g. taking or not taking into account indirect references, taking or not taking into account creationists’ works, etc.)

The good thing about this approach is that it is not too hard to do it: We could value prediction markets of prediction markets just by the usual crypto voting procedure. This voting can be done (among other) in the following ways:

  • one person – one vote (needs an anti-Sybil system such as BrightID [7])
  • any kind of DAO (e.g. an Aragon [8] DAO)
    • quadratic voting (also needs anti-Sybil) for Aragon is not yet done well enough)

The Mathematical Model

For a visual summary of the below see figure 1.

The following is a mathematical model (applicable to science financing but not only) possible to implement in blockchain for stimulating market to implement the author and publisher scoring oracles.

  1. Using a special smart contract anybody can claim his account (just call a special contract method with his Ethereum address and probably some identity like first/last name) as a science author for purposes of his “salary” accounting. Simply, due to technical limitations of ERC-1155, a user needs to be explicitly registered before he can receive tokens.
  2. Likewise, using a special smart contract (ditto) anybody can claim his account as a science publisher.
  3. Anybody could create (if he has enough money to implement this) an oracle (a possibly off-chain software that may write data to the chain using a specially provided method) mapping for example scientist’s Ethereum address to the scientist’s score (such as based on the number of direct and indirect citations and on how free the manuscript license is) (accordingly this oracle) and optionally to the publisher(s). That is oracle is a software that stores into the chain scientists “score” (a number which his salary is proportional to, calculated accordingly the oracle’s creator’s sense of fairness). With an oracle is also associated an address that will hold collateral funds for this oracle (after 100 years), to be further distributed among scientists.
  4. Create two prediction markets:
    1. the score of each oracle
    2. the score of each scientist and each publisher by each oracle (so NxM where N is the number of scientists and M is the number of oracles outcomes)
  5. Mint to each scientist his own conditional token proportional to time passed since his registration. This will be his salary.
  6. Allow anybody to deposit a token to be exchanged for the collateral (after 100+ years, see below).
  7. After 100 years each oracle owner is expected to write the “score” (presumably based on citations counts and licensing/pricing policies) his conditional token for each scientist (with nonzero results) into the blockchain. He is also expected to score publishers. (A probably simplest way to score publishers is to add scores for all articles for which it is is considered the primary publisher.)
  8. During some additional time period the people (or robots) vote resulting scoring each of these oracles 0..1 (with the sum 1) based on how they perceive the “correctness” or “fairness” of its scientists’ scores (for our purposes we define correctness/fairness of an oracle as the quantity of votes an oracle will receive). The voting results are the outcomes of the oracles list predictions.
  9. Each oracle receives the collateral proportional to its score.
  10. Each conditional token holder can receive (after 100 years, for any oracle) the collateral proportional to the product of his score in an oracle to the score (voting result) of the oracle using two consequtive transfers: 1. to oracle’s collateral holder from the “prediction of predictions” contract; 2. from that collateral holder to the requestor of funds. (To be retrieved from each oracle funds separately, because potentially there may be many oracles.)

We have thus a prediction market that will run for about 100 years. As I explain in this note, this kind of prediction markets is similar to transferring money from the future. So, we have the prediction markets for scientists, publishers, and oracles, they receive some reward now (not after 100 years). How to split the collateral into three parts between scientists+publishers and oracles? It could be done by voting after 100 years with voting results taking the arithmetic averages from each voter. Or we can start voting now and keep it running for 100 years. We could also pay to reviewers but that’s not possible to do fairly as they are anonymous.

The incentives to put a collateral into this market are:

  • goodwill to help the scientists, etc.
  • the donor will receive more voting rights.

The market resolves only after 100 years, but as we know conditional tokens have value even before the market resolves. It incentivizes to trade the tokens meantime.

Multi-token system

The salary sometimes needs to be recalculated (for example, when a person dies, it is probably reasonable to set its exchange rate to zero). So we allow anyone to call a special method to recalculate anybody’s other’s (or his own) salary. In this case a new token is created that replaces the old salary token.

Another reason for producing new salary tokens sometimes is to prevent traders to profit from killing their clients (just like as someone may kill a bequestor bequesting to him).

To make trading convenient and efficient, it is possible to create a wrapper token contract that “combines” several tokens into one.

Can a good scientist earn $100000M per year soon?!

5: Wolfers, Justin, and Eric Zitzewitz, Prediction Markets, 2004

6: Gnosis Ltd., Gnosis, , https://gnosis.io