Token Engineering Canvas & Agent Behaviour Map + basics for Token Engineering

When it comes to engineering token incentives we are still at the early days of learning how an approach with engineering in mind will lead to reach clarified designs in less time.

For those new to the initiative of #tokenengineering, please visit, read and subscribe at http://tokenengineering.net/ (the Wiki).

As part of this community driven initiative we (me and Dimitri De Jonghe) would like to present the first version of the token engineering canvas that is designed as a “cheat sheet” for the design. The canvas is there to help token-entrepreneurs and support token-economist to have a simple framework to follow while designing incentives for their platform. The overall goal with the cheat-sheet is to organise thoughts in a structured way with a generalised step-by-step guide. Special thanks for the #tokenengineering gorup for the mind opening conversations and the review of the draft of this model.

The creation of the canvas was inspired by hands-on experience from the Berlin chapter of the TE community. The below article is the aftermath of the meetup where Taqanu was in crosshair from a TE perspective lead by Trent McConaghy.

For the first time engineering approach was applied at this meetup. This was the first TE meetup with focus on the engineering side following the successful kickstart of the movement during Consensus in NYC.

This Token Engineering Canvas is not a tokenized translation of the Business Model Canvas by Alexander Osterwalder as it is very specifically designed to support the creation of token design thus it required us to rethink the process and requirement for an engineering like approach. It borrows the word Canvas due to its wide spread recognition of the BMC, however, the differences between a business model vs an engineering driven token design does not enable us to draw direct correlations.

Token Engineering Canvas

The use of the canvas

First and foremost, the use of the TEC does require some level of knowledge about blockchain technology as it requires engineering knowledge. As a non developer myself I can assure people with interest in the space that its possible to use it with ease, but for someone unfamiliar with the space it will not provide much aid.

The use of the canvas is meant to be very straightforward.

1, Fill in basic informations for traceability. Project name, the names of the collaborators and the date.

2, Determine the overall objective of the project (eg. Bitcoin — maximise security). Do not hurry with this step as misunderstanding the goal of the project might lead to faulty design.

3, Start to set objectives and constrains that has to be reviewed during the process. Be as specific as possible.

4, Start the iteration process from the most simple pattern design. (eg. Proof of Work without longest chain)

5, In case non of the formerly known models fulfil the requirements start the development of a new pattern that delivers on the preset requirements.

The Token Engineering Canvas can be downloaded and commented HERE.

The TEC has a second page that enables the engineering approach to unfold in the hand of the user called Agent Behaviour Map or ABP.

Agent Behaviour Map (ABP)

It is important to remember that incentive mechanisms only work with agents participating in the predetermined framework. Without participation the system stops being operational. Agents are considered as inputs. It is important to have a complete understanding of expected and possible behaviours from any possible input source to maintain confidence in the system design.

Agents should be categorised by general and measured from a variety of perspectives with an additional deterministic approach using BAR (Byzantine, Altruistic, Rational) /Aiyer,2005/ models for each possible agent.

Including the AMP as part of the TMC was important as it gives the depth of understanding that is required to design a working system.

By adding this level of complexity to the TEC we move the model from a two dimensional canvas to a three dimensional map. The three axes are:

  • objectives and constraints
  • the pattern design(s)
  • the participating agent type(s)
Agent Behaviour Map

The same patterns (represented by numbers) are used across both pages.

Simulation — BitCoin:

We included a first simulation for BitCoin on the second part of the Canvas so it can be used as an example for other projects.

If you have additional projects analysed please add it in a new tab for review by the community. In case there will be too big interest to list projects we will think how we could make it readable.

Token Engineering Canvas used for analyzing BitCoin

Useful links:

Both the Canvas and the Map includes the general fill out steps so you will not need to constantly keep this post open. For reference and to comprehend the level of complexity I would strongly recommend to read Trent McConaghy’s post about the entire practice of TE to have a good understanding what it is about:

Other useful sheets for mapping out possible patterns for the canvas:

source from Dimitri De Jonghe: https://docs.google.com/presentation/d/1CqMs85_0MZ6zzyWXigBKInMbEC6VcMB-4O46J_l-3W4/edit#slide=id.g24e133c43f_0_0

And a clear classification of tokens by Paratii using previous deep dives by Balaji S. Srinivasan - thoughts on tokens and Kyle Samani - new-models-utility-tokens

original source from Felipe Gaúcho Pereira : https://medium.com/paratii/on-the-immaturity-of-tokenized-value-capture-mechanisms-1fde33f2bc8e

To better understand each category of tokens read Felipe’s medium post on “On the immaturity of tokenized value capture mechanisms” where he explains not only the intricacy of token types but also the general notion of the economic powers of token mechanics.

Thank you Trent McConaghy, Angela Kreitenweis, Dimitri De Jonghe for inviting me to the working group of http://tokenengineering.net/.

This is a work-in-progress document and anyone with knowledge and experience over the topic is invited to contribute.