Impermanent loss has become a much-criticized problem for AMM. During the rapid development of defi, a number of technical solutions have emerged to solve this problem, such as options represented by Opyn, insurance solutions represented by Instaraise, and bancor by BNT compensation for IL. Recently we found that some projects try to solve the impermanent loss problem through the oracle. However, after careful research, we found that these schemes do not have a deep understanding of impermanence loss. The essential reason of impermanence loss is that the AMM pool must sell low buy high to provide liquidity during the process of price movement. Can sell high buy low be achieved through the oracle machine? Of course the answer is No. For example, the following picture is the test data of some project.
The test data shows that the pool did not incur much IL during A and C. They think that the pool was mainly selling SOL because the price of oracle, which mainly comes from CEXs and traders (via Pyth), always moves faster than constant product AMMs, so it was sold in advance. However, as we all know, the oracle is not a forecasting machine, any oracle is a reaction of historical prices, and there is no inevitable causal relationship between future market trends and the oracle. It can be that cex leads dex, or dex leads cex. No one can be a prophet. If they can, then the market will lose the meaning of existence.