4 ms·
Nope. Depends on what kind of security you are looking for. Which door is more likely to reliably allow me to enter my home? 1. No door. 2. A door with N lock
by RichardHeart 9y ago
Nope. Depends on what kind of security you are looking for. Which door is more likely to reliably allow me to enter my home? 1. No door. 2. A door with N locks in parallel. 3. A door with N locks in series. As N increases, or goes serial instead of parallel, my ability to enter my own home reliably decreases. Consequently, assuming so side channel attacks (windows or other doors) the chance that other people can't get in also increases.
Understanding that adding variables in serial in order to fuel your car reduces the chance you can successfully fuel your car is easy. Understanding that when the internet goes down, you won't get gas is easy to understand.
Hedges aren't what they're cracked up to be. 1. What happens when your hedge counterpart fails? (check out credit default swaps, or trades that can't execute, or limit down markets, etc etc, bad oracles (credit rating agencies..)) All hedges require a counterparty. Counterparties fail. Even when they don't fail, they become corrupted by bad oracles.
You can't create reliability by wiring variables in series. You can create "security" at the cost of reliablity. How much cash has been lost in "smart" contracts, vs stupid safes per capita.
- darawk 9y agoNone of that is correct. 'Series' or 'parallel' are the wrong way to think about things. If you buy 100 shares of AAPL and then you buy a put on SPY, that position is objectively lower risk than the unhedged AAPL position. Your point about counterparties is irrelevant, as I didn't claim that a hedge causes risk to go to zero - merely that it reduces it. It is possible for two risky things to cancel one another (or some components of one another, as in the AAPL/SPY example, the SPY put hedges AAPL's beta).