Thread by @MaxResnick1

Does your proof of stake blockchain have weak pullout game? In our latest paper, @mikeneuder, @mallesh, and I share a few tricks to help you reduce wait times while achieving the same security when designing exit queues. 1/9

🧵👇

Designing exit queues is all about balancing security and flexibility. If you let too many validators exit within a short period then accountability guarantees go out the window, but locking validator stake in for too long is bad for staker welfare, meaning the protocol has to give out more, in the form of emissions, to compensate. 2/9

Formally, the blockchain designer in our model faces potentially multiple security constraints of the form: “no more than x% of the stake can exit within a time window t”. The goal is to produce a mechanism that satisfies these constraints but minimizes disutility from waiting. 3/9

When stakers are homogenous in time sensitivity, a simple dynamic capacity first come first serve queue is optimal. 4/9

In each period, the mechanism processes exits in the order that they arrived, but processes as many exits as possible without violating any of the constraints. This means exits are processed much faster than in a static capacity first come first serve queue. 5/9

When stakers have heterogenous time sensitivity, the optimal mechanism must allow more time sensitive users to cut in front of less time sensitive users to achieve efficiency. The optimal mechanism here is a dynamic capacity priority queue. 6/9

We also discuss what protocols use in practice in the last section! 7/9

Read the full paper here! mechanism.org/spec/06

Also available on arxiv! arxiv.org/abs/2406.05124

oops wrong tag @malleshpai