Thread by @thorchain_org

For a system to secure assets that are external to it, the economic cost to attack the system must always exceed that value of assets it is securing.

If there ever is a case where more can be gained from an attack than it costs, the assets are insecure.

Proof-of-Authority networks (such as @Blockstream Liquid) select for entities to put their reputations risk.

These networks are more difficult to assess, since actors are also assumed to be partially altruistic (eg, Blockstream), and reputation costs are subjective.

What is desired is a system that can be built purely on game theory assumptions; that all participants are rational profit-seeking actors.

This narrows the assumptions and the system’s security can be objectively measured.

Since THORChain is a proof-of-stake network, the costs to attack are purely capital-based, so a system of incentives can be designed to target the desired behaviour.

To ensure assets are secure, the system must become aware of the value of the assets it is securing compared to the asset which provides security.

Thus a price oracle is needed to transmit the purchasing power of both assets.

Price oracles are an unsolved problem and are notorious for being central points of failure.

The only trustless source of purchasing power is a liquidity pool that binds the two assets together.

However, liquidity pools with fixed-rate fees are vulnerable to sandwich attacks, so the solution is to use slip-based fees.

By charging slip-based fees, sandwich attacks become prohibitively expensive.

THORChain also uses its liquidity pools as a mechanism to solve liquid asset exchange, the very purpose of its being.

Liquidity pools solve for incentivisation, availability and redemption speed.

This system is so elegant that in time all other networks will converge to it.

THORChain's design is terminal. It *IS* the only design that will work. $RUNE

It's probably time for @zhusu to weigh in with his updated opinion as well.