Since Satoshi did not solve the Byzantine generals problem, this means confirmations are completely arbitrary. So why are two confirmations more useful in Bitcoin (PoW) than one? Because it's an open entropy system where over a period of time, it's either unlikely or statistically impossible for someone to maintain a monopoly on block validation when there is no upper limit to confirmations.
Recursive systems like proof of stake tend to permanently monopolize block validation by design, with no real fault or state recovery to fix it once it goes off the rails. This makes a proof of stake confirmation essentially worthless due to being a bounded entropy system.
On top of being worthless, proof of stake is also a permissioned ledger. The purpose of mining in Bitcoin is to create a permanent decentralized exchange peg, which thus results in a permissionless system.
r0ach, I admire your patience trying to teach thermodynamics to presumed trolls
Especially how most - even well intended fellows here - do not get why PoS is an utter failure as a network security model (added, the economic incentive model is counterproductive, resulting in very hackable, centralized like security!).
Just imagine their horror as we would explain them how the early miners effectively would still be able to control ~10-20% of the network security with their laptop made coins hodled in a PoS wallet in the era of huge asics farms Hillarious!