We began on a journey looking at consensus mechanisms starting at Proof of Authority and Proof of Stake. We started with questions of what makes these mechanisms worked and published our early findings. The Interchain Foundation awarded us a grant to further research our ideas for Proof of Engagement and develop simulations.
The research paper and the simulations have been published and we have drawn some interesting insights from the research and the preliminary modeling results are encouraging. Clearly there is further modeling which can be done to further test the scenarios, and at this stage we see stability of the results to changes in the initial parameters, at least within some range, which is a good metric to see how well this model generalizes.
Proof of Engagement provides the incentives for all of the community around a blockchain to be actively involved.
We think measurable incentives to COLLABORATE are essential for a sustainable decentralization and Proof of Engagement provides the framework for this to happen.
Proof of Engagement brings higher decentralisation through the combination of Engagement Rewards and Stake than a pure Proof of Stake mechanism.
Proof of Stake makes it is easy for validators with mid/high stake to capture voting weight by offering low commission.
With Proof of Engagement there is little scope for validators to capture voting weight unless, in some cases, they have a relatively high level of engagement.
Yield-seeking delegators push voting weights in the direction of more equality and make profits slightly less equal across validators.
Most validators have lower profits after delegation. High-stake, low-engagement validators lose the most profit, and the biggest winners are validators with low stake but high engagement.
These findings point to a more decentralised, and robust consensus mechanism which make a good foundation for a secure blockchain.
The full paper, models and simulation results can be found here:
Thanks to the team at Confio, Ethan for co-authoring and developing the ideas, Will for writing the simulations. Additional thanks to Dr Moustapha Awada, Dr Chulwoo Han for reviewing, and for the many people we discussed the ideas with.