Vitalik explains in his essay Why proof of stake that proof-of-stake is more secure and more decentralized than proof of work.
(1) Proof-of-stake systems are more expensive to attack. In summary, pow systems are mostly hardware costs whereas pos systems are mostly capital costs.
- Vitalik theorizes a small calculation to show that it’s more expensive to be a malicious actor on a proof-of-stake network than on a proof-of-work network.
- The total cost of attack on a GPU-based proof of work system is 486.75
- From Ethereum.org, Ethereum was mostly mined using GPUs rather than ASICs, which kept the cost down (although had Ethereum stayed on proof-of-work, ASIC mining may have become more popular)
- Per Vitalik, GPUs are relatively inexpensive to rent
- The total cost of attack on a GPU-based proof of work system is 486.75
- Attacks in a proof-of-stake system are much easier to recover from than in a proof-of-work system.
- In a pow system, an attacker can make a chain permanently useless with a spawn camping attack. Honest miners will drop out. If community implements a hard fork, miners on the attacked fork would be “bricked” making them useless.
- On the other hand, a proof-of-work system has a built in lashing mechanism where only the bad actor, and no one else on the network, has their stake destroyed. And for or harder-to-detect attacks such as a 51% coalition censoring nodes on the network, the community can coordinate a minority user-activated soft that destroys the attacker’s stake (2) Proof-of-stake systems are more decentralized iffy
- PoW is a hardware arms race that tends to price out individuals and small entities. If Ethereum stayed on proof of stake, it would have probably started to use ASIC mining
- Interesting because proof-of-stake algo I believe favors those with more locked capital?
Other benefits: Electricity consumption issue was definitely big in the news and a big deal.