The 2018 paper “On profitability of egocentric mining” by Cyril Grunspan and Ricardo Pérez-Marco [0] introduces a selfish-mining assault that depends on the problem adjustment algorithm. Their followup paper (similar 12 months) “On profitability of cussed mining” [2] extrapolates this to assault variants.
It first explains how, when issue is fixed, selfish-mining is much less worthwhile than sincere mining (see additionally [1]). However then it claims that the problem adjustment algorithm introduces a method to make revenue. Earlier than I even try to grasp the maths, I am having a tough time making sense of this intuitively.
By quickly withholding blocks in interval N the attacker is growing the stale block price, each for themselves and for sincere miners. This causes a drop in issue which makes mining within the subsequent interval extra worthwhile. Up to now I comply with.
However why would a decrease issue in interval N + 1 profit the egocentric pool extra than its rivals? It appears to me all swimming pools revenue equally in interval N + 1. However the egocentric pool needed to make an funding in interval N, since we have already established that at fixed issue this technique is much less worthwhile.
In order a impartial miner selecting between swimming pools, you do not need to be within the attacker pool in interval N when it is dropping cash. You additionally do not need to be part of it in interval N + 1 as a result of you’ll be able to reap the identical reward in every other pool.
So are they making an assumption about pool loyalty?
[0] https://arxiv.org/abs/1805.08281