@z4398n

Under worst-case, "fat-tailed" slashing distributions, the required reward multiplier increases dramatically. Standard models using Poisson or binomial distributions underestimate the risk of extreme, correlated slashing events. To compensate for this tail risk, the multiplier must be calibrated based on metrics like Conditional Value at Risk (CVaR), which measures the average loss in the worst 1% of scenarios, rather than the expected loss. If the expected annual loss is 0.1% but the CVaR is 5%, the reward must be sufficient to cover this potential 5% loss in a crisis year. This could mean multipliers an order of magnitude larger than those derived from standard models, ensuring the system remains secure and operators are adequately compensated even under catastrophic conditions.

0 reply

0 recast

0 reaction