The premise centers on the probability of producing adequate #s of paired but opposing outlier component positions which may earn more on gamma than is lost to the premium paid on the vol-swap, component vol // index vol. Paired-opposing outlier positions will be offset and replaced with fresh atm gamma.
Wonder if arb would call somebody that right to their face?