Today the Ethereum Foundation has published a post about Ethereum 2.0.that explains the possible scenarios, related to incentives, in case some actors should become threats to the protocol causing it to be unusable.
A tool that has been developed to prevent this problem is called slashing, which has two different functions:
To prohibit attacks against ETH 2.0;
Preventing validators from being lazy.
In fact, slashing destroys part of the crypto funds staked by the validator who decides to behave negatively or doesn’t vote for the system or even votes several times – the so-called double voting. The aim is therefore to achieve network uniformity with regard to the blocks that need to be validated.
These are the only reasons for a validator to be put out of phase 0, but new measures will be introduced so that validators store the information and shards as well as make the latter work.
As far as the destruction of the validators’ stake is concerned, this is done in a proportional way depending on how the validators attacked Ethereum since it takes a large number of validators to carry out the attack. Whoever is affected by slashing will be removed from the blockchain and will no longer be able to participate.
Another case analyzed by the Ethereum team is the one in which a validator would be offline: in this case, they would lose the gains they should have received in that time, considering that the more time a validator is online, the greater the amount of crypto they are awarded.
Another interesting aspect mentioned in this post about Ethereum 2.0 scenarios is the case when more than 1/3 of validators go offline. In this case, the system is designed to keep running anyway, eliminating offline validators and, once 2/3 of the staking power is restored, then the protocol would resume the normal course.
Again, when offline validators are deleted, they would lose part of their funds.
A different story, however, with regard to the rewards that are given to the validators: both a fixed share of the reward and one based on inflation will be used, so they will have both a fixed share to do their job and a variable share that will depend on the number of validators that will be involved.
In this way, it is possible to incentivize validators to work within the system without receiving ridiculous rewards.
Finally, the post ends with a remark to keep in mind: “hope for the best, but expect the worst”.