Further discussion of Making EIP 1559 more like an AMM curve

[barnabe]

@llllvvuu

waiting time

Related to this, the distance between the clearing price and the true market price could also be a metric of good UX. e.g., if the market price is 1, but the clearing price keeps oscillating between 0 and 2, I would call that bad UX. We investigated this metric in simulations here.

@MaxResnick

I do wonder if it is more natural to treat the burst constraint as a threshold i.e. block size is less than some threshold rather than a more complicated one, then we can come up with a nice estimate for what the state growth cost is and use that to define the curve.

Separating burst and running constraints does allow for such optimisations, e.g., you can price state growth much more precisely along some network harm curve vs burst resources which are either no problem load-wise until they cross a threshold and become a problem. The post by @vbuterin on Multi-dimensional 1559 goes over this.