Inspiration
With Ethereum moving from a PoW (Proof-of-Work) consensus protocol to a PoS (Proof of-Stake) one via CASPER, we thought it appropriate to analyze how the governance model of PoS fit in with the Ethereum ecosystem and what kind of potential advantages and threats came with the new protocol. After talking to Karl Floersch about his intended design for CASPER, we came to the conclusion that a good PoS system has yet to be implemented and that CASPER, being a pioneer in PoS, still had some holes that needed to be filled. So we took it upon ourselves to look at how we might help CASPER by looking at different attack vectors and propose a potential ideal block reward amount.
What it does
Our paper talks about the current intended implementation of CASPER and goes on to dive into some possible threats that might occur and how we might incentivize people into acting honestly based on block reward.
How we built it
We used a Markov Chain Model to predict how a stake holder with large initial stake might be able to reach an aggregated 33% stake value over time. We also used financial modelling, comparing the block reward from being a validator in CASPER to potentially investing in more traditional financial assets (such as risk-free assets and high risk/high reward assets) as a means of circumventing the network and aggregating 33% of stake value.
Challenges we ran into meh
Initially we ran into problems on how to quantify the Markov Chain Model given that the probabilities of any event were changing at regular intervals. We also ran into problems where we just didn't know what CASPER intended on doing, and so, had difficulties in making our model when it came to what assumptions were safe to make and what comparisons in the market were appropriate and bounded.
Accomplishments that we're proud of
Being able to contribute to an ACTUAL consensus protocol that might actually be used as part of the Ethereum network in itself is an amazing feeling. Moreover, after talking to Karl, we realized that they don't have a solid plan on how they intend on implementing their reward scheme for miners and validators in the network. So, being able to contribute some kind of research where we can help them find an ideal reward scheme is something we all enjoyed doing in and of itself.
What we learned
We learned that PoS has the potential to be a better consensus algorithm than PoW, but there are still MANY bugs that need to be worked out.
What's next for Parameterizing Casper
Next, we intend on using better mathematical modelling to see how different reward schemes would work. From that, we would be able to give a tighter interval on what an ideal reward scheme might look like for CASPER
Log in or sign up for Devpost to join the conversation.