UPDATE:
Here’s what I think has been missed by the PBS designers…
PBS does what it claims if the dominant builder is not a pool, but it fails catastrophically if it is (ie: does the opposite and worsens centralization).
Trouble is, pools want to be the dominant builder because it increases their market share over competing pools- and they have the resources to do so and latency advantages to help them.
You can’t deny this is their aim, because it is this that PBS was designed to address.
So it is far from clear that PBS will have a positive impact on validator decentralization.
(To model the scenario where MEV is not extracted by the pools, which is what Flashbots hope will happen, you can use these parameters to add a dominant mev extractor that is not a pool.
poolStartingAllocations = new double[] { 0.3, 0.15, 0.14, 0.09, 0.06, 0.03, 0.23, 0.0001 };
poolMevEfficiency = new double[] { 0.7, 0.7, 0.7, 0.7, 0.7, 0.7, 0.7, 1 };)