## Abstract

This article presents the initial results of the first governing body based on the principles of Fractal Governance[1]. 

Initial results are presented from an analysis of the dataset from the first social experiment in Fractal Governance. This dataset and associated tooling are made available in an [open source GitHub repository](https://github.com/matt-langston/fractal_governance) for data scientists, researchers and educators to explore the practical application of Fractal Governance to organize people in the production of public goods and services for mutual aid.

## Introduction

The goal of Fractal Governance is lofty: incentivize people to collaborate in the production of public goods and services that also eliminates corruption and graft in the process.

The principles of Fractal Governance are described in the book [More Equal Animals](https://moreequalanimals.com) by Daniel Larimer, and the technical specifications for how to implement Fractal Governance at scale is defined in the [fractally White Paper](https://fractally.com). This article presents an analysis of the initial results of the first group of people to govern themselves according to these principles and technical specifications.

These pioneers call their group Genesis and refer to it as a Fractal. The Genesis members meet weekly to mine the inherent value of their collaboration to produce public goods and services and return that mined value, tokenized in units called *Respect*, directly back to its members through a governance process that naturally prevents corruption and graft. This incorruptibility is the defining feature of Fractal Governance.

Fractal Governance directly and consistently rewards individuals for their recent past contributions towards the creation of public goods and services that also avoids the formation of Pareto distributions due to corruption and graft found in all other known forms of governance. Gone are the days of rewarding collusion with illicit gains (such as currency) from dishonest behavior or other questionable means.

Analogous to Bitcoin's *Proof of Work* consensus algorithm which rewards people for transforming stored energy, in the form of electricity, into an incorruptible public ledger of account, a collaboration of people governed in a Fractal nature also uses a *Proof of Work* consensus algorithm to reward people for transforming stored energy, in the form of human collaboration, into public goods and services.

The fundamental difference between the two consensus algorithms is in how rewards are allocated. The Bitcoin model allocates rewards called BTC tokens to those who consume the most electricity most consistently. The Fractal model, on the other hand, allocates rewards called Respect tokens to those who contribute the most value most consistently, as judged by their peers.

The Bitcoin consensus algorithm is prone to corruption and graft because it rewards those that obtain the most consistent source of electricity by any means whatsoever, illicit or otherwise. The Fractal Governance consensus algorithm, on the other hand, prevents corruption and graft by eliminating opportunities for collusion.

The nature of how the rewards from the Bitcoin and Fractal Governance systems are recorded is similar in  that both systems use a blockchain for their public ledger of account.

## Experimental Apparatus

The experimental apparatus for the results presented in this article are the weekly consensus meetings of the Genesis membership that are coordinated via a [dedicated web portal](https://gofractally.com/groups/7064857/feed).

The first weekly consensus meeting was held on February 26, 2022, and as of the date of this article there have been 11 weekly consensus meetings, the most recent of which was held on May 14, 2022.

During each weekly consensus meeting, which lasts about an hour, members are randomly assigned to small groups of five or six people and tasked with evaluating each others recent past contributions by judging one another's answer to this question:

> What have you done to build the technology, raise awareness and grow adoption of fractal governance, as per fractally's white paper.

A member's answer to this question is the *public good and service* currently valued by the Genesis Fractal because understanding the true nature of Fractal Governance is required to prove whether it is an incorruptible alternative to other forms of governance that distribute public value according to Pareto distributions through corruption and graft.

Understanding the true nature of Fractal Governance is through experimentation, and two types of experiments are being pursued:

1. Social experiments, as described in this current article.
2. Simulation experiments, which is the subject of a future article.

The value of each member's contribution is evaluated and force ranked by the consensus of their small group and the results are recorded on a blockchain. The Hive blockchain is currently used to record each small group's consensus while a purpose-built replacement is in development.

The data for this current article is curated on a weekly basis by [Gregory Wexler](https://gofractally.com/members/10362727), [Joshua Seymour](https://gofractally.com/members/10361546) and [Matt Langston](https://gofractally.com/members/10426315).

## Results

The results presented in this section are from an analysis of a [curated  dataset](https://raw.githubusercontent.com/matt-langston/fractal_governance/main/data/genesis-weekly_rank.csv)  from the first 11 weekly consensus meetings of the Genesis Fractal.  The dataset includes all meetings up through May 14, 2022.

This dataset records the weekly mining activity of the Genesis membership, and both it and the associated data analysis tools[3] are made publicly available in an [open source GitHub repository](https://github.com/matt-langston/fractal_governance).

### Summary Statistics

|Metric|Measurement|
|:---|---:|
|Total number of unique participants|97|
|Total Respect tokens earned from all sources|4,050|
|Total Respect tokens earned by members|3,317|
|Total Respect tokens earned by teams|733|
|Average number of attendees per meeting|33 +/- 9|
|Average number of meetings attended by a unique member|4 +/ 3|
|Average team representation per meeting[2]|0.22 +/- 0.07|

### Observations and Discussion

The attendance over time remains consistent with 33 participants on average attending the weekly consensus meetings. This is remarkably stable given that any particular participant only attends four meetings on average. The following two plots help visualize these observations.

![Attendance vs Time](https://github.com/matt-langston/fractal_governance/raw/main/README_files/README_26_0.png)

![Consistency of Attendance](https://github.com/matt-langston/fractal_governance/raw/main/README_files/README_31_0.png)

This phenomena of consistent stability in overall participation is due to a *core group* of dedicated members juxtaposed with a steady stream of new participants as knowledge about Fractal Governance becomes more widely understood.

The members of the *core group* consistently contribute value and rarely miss a meeting. The *wisdom of the crowds* consistently identifies these top performing individuals and rewards  them handsomely with Respect tokens.

However, it is not just these top performers that are awarded with Respect tokens. Everyone who participates in a weekly consensus meeting is rewarded with some amount of Respect tokens, even if they have done nothing else other than participate in a meeting. This is because meeting participation is how community value is identified or *mined*  via the *wisdom of the crowds*, and is how Respect tokens are minted into existence; this is the equivalent of Bitcoin miners receiving newly minted BTC tokens for their mining services.

The following plot helps visualize the mining and minting of membership Respect tokens over time.

![Accumulated Member Respect vs Time](https://github.com/matt-langston/fractal_governance/raw/main/README_files/README_44_0.png)

The mean accumulated consensus rank has a correlation coefficient of 0.6 with meeting attendance. Based on this strong correlation we perform a linear regression and overlay the results on the following plot of rank vs the number of meetings attended.

![Attendance Count vs Rank](https://github.com/matt-langston/fractal_governance/raw/main/README_files/README_72_0.png)

As the plot demonstrates, members rank higher on average in subsequent weeks based on the number of past weekly consensus meetings they have participated in. Possible reasons for this phenomena include:

* Over time members learn what their fellow members value and come into alignment with those values.
* Over time members begin to imitate their higher ranked colleagues from watching how they conduct themselves.
* There is a self-selection process going on. This is an interesting idea for further analysis.

#### Teams
An important element of Fractal Governance is the formation of teams. Teams are voluntary associations of individuals who have agreed to combine their efforts and work together towards creating a public good or service for the benefit of their Fractal. Because teams create value for a Fractal beyond what any individual can create on their own, a team is awarded Respect tokens in amounts equal to the Respect tokens awarded to its individual members during their participation in the weekly consensus meetings.

For example, if a team member earned 21 Respect tokens during a weekly consensus meeting, then 21 additional Respect tokens are minted and awarded to the team, for a total of 42 newly minted Respect tokens.

The internal governance of a team is left unspecified, other than a minimum tenure for its members before they can leave one team to join another. Therefore, each team decides what to do with their team Respect tokens, which could simply involve distributing them back to individual team members. 

There are currently three teams that on average compose 22% of attendees of the weekly consensus meetings. The following two plots helps visualize the mining and minting of Team Respect tokens over time, as well as their representation during the weekly consensus meetings.

![Accumulated Team Respect vs Time](https://github.com/matt-langston/fractal_governance/raw/main/README_files/README_53_0.png)

![Team Representation vs Time](https://github.com/matt-langston/fractal_governance/raw/main/README_files/README_58_0.png)

## Conclusion

The initial results of the first governing body based on the principles of Fractal Governance was analyzed and presented. This governing body was formed on February 26, 2022 and is called the Genesis Fractal. During its first 11 weekly consensus meetings there have been 97 unique members participate with an average of 33 participants per meeting. Three teams have formed during this time and 4,050 Respect tokens have been minted.

The results presented in this article are also published to a live [Genesis Fractal Dashboard](https://share.streamlit.io/matt-langston/fractal_governance/main/fractal_governance/streamlit/genesis_fractal.py)  that updates automatically as new weekly results are curated and published.

## Resources

Resources to learn more about Fractal Governance:

- [fractally White Paper](https://fractally.com)
- [More Equal Animals](https://moreequalanimals.com) by Daniel Larimer
- [Genesis Fractal Dashboard](https://share.streamlit.io/matt-langston/fractal_governance/main/fractal_governance/streamlit/genesis_fractal.py)
- [Modeling and Simulation](https://gofractally.com/groups/7064857/topics/7623063) topic on [gofractally.com](https://gofractally.com)


[1]: The word *fractal* is used throughput this article and is capitalized as *Fractal* depending on the context to distinguish between the usual mathematical meaning and its specific use when describing its application to governance.

[2]: The "average team representation" is the number of members in attendance that are members of a team divided by the total number of members in attendance.

[3]: A Jupyter notebook, Streamlit dashboard and multiple pandas DataFrames provided by a python library.