Now I’m well aware that Ethereum is migrating to Verkle trees which are a completely different beast, but while working on my own blockchain project I evaluated Merkle-Patricia trees (or tries) versus various types of self-balancing trees and I suspect I have a result that makes 3-ary B-trees significantly better than MPTs. Let’s consider a highly simplistic scenario where we only need to index account balances, so no smartcontract storage or anything else. We need to construct a Merkle-p...
