> Weighted users.To prevent Sybil attacks, Algorand assigns a weight to each user. BA⋆is designed to guarantee consensus as long as a weighted fraction (a constant greater than 2/3) of the users are honest. In Algorand, we weigh users based on the money in their account.
That 2/3 is your n=3f+1 right there. It's just that, as I said of PoS in general, the n is based on coin holdings.
This is not controversial, as the authors of Algorand are completely aware of the limitation:
> Most Byzantine consensus protocols require more than 2/3 of servers to be honest, and Algorand’s BA⋆ inherits this limitation (in the form of 2/3 of the money being held by honest users). BFT2F [35] shows that it is possible to achieve “fork∗-consensus” with just over half of the servers being honest, but fork∗-consensus would allow an adversary to double-spend on the two forked blockchains, which Algorand avoids.
If you read the whole paper you will see that it scales to more than the 3f+1 participants by randomly selecting a committee (of 3f+1 validators) at each round.
> Weighted users.To prevent Sybil attacks, Algorand assigns a weight to each user. BA⋆is designed to guarantee consensus as long as a weighted fraction (a constant greater than 2/3) of the users are honest. In Algorand, we weigh users based on the money in their account.
That 2/3 is your n=3f+1 right there. It's just that, as I said of PoS in general, the n is based on coin holdings.
This is not controversial, as the authors of Algorand are completely aware of the limitation:
> Most Byzantine consensus protocols require more than 2/3 of servers to be honest, and Algorand’s BA⋆ inherits this limitation (in the form of 2/3 of the money being held by honest users). BFT2F [35] shows that it is possible to achieve “fork∗-consensus” with just over half of the servers being honest, but fork∗-consensus would allow an adversary to double-spend on the two forked blockchains, which Algorand avoids.