Re-staking / shared security system, to put it simply, is "borrowing multiple times against the same collateral." The returns may seem to stack, but so do the risks—it's just that you don't see it on the dashboard. Especially with those "reduction conditions / trigger windows / correlation assumptions" in the contract parameters—changing them means it's no longer the same thing. Don't just focus on the annualized line.

Recently, some people have been criticizing that on-chain data tools and tagging systems are lagging and can be misleading. I believe: if you use unreliable attribution to calculate risk exposure, you'll end up taking "appearing diversified" as truly diversified. Anyway, when I look at staking projects now, I first ask: who is actually backing the guarantee when things go wrong, and will the backing assets all fail at the same time? If you can't see clearly, it's better to avoid it—no need to gamble with illusions.