Research Questions:

• How to normalize stake values in crypsinous?
  • why do we need epochs?
    • if we can get rid of epochs, we can get rid of stake freezing and use total supply to normalize the stake values.
  • Can we learn the normalizing term by counting empty slots

Thoughts

• Stake does not necessarily need to be locked, simply owning the chain currency is proof enough that you have something at stake. Since traditional PBFT protocols are more affected by large validator churn, they want to lock stake to reduce turnover. We don’t need this with Ouroboros
  • but we may end up requiring it with crypsinous.

Problems that remain to be solved:

• Leader Election probablity distribution assumes knowledge of total stake:

  

  We don’t have this information in crypsinous, thus we need to estimate it somehow

Classic

Praos