I was digging through @Dusk ’s consensus material and kept coming back to one detail: committee selection is deterministic, but selection probability is tied to stake. It looks like an implementation choice. I think it is more fundamental.
Dusk’s documentation says deterministic sortition computes a score from consensus-round parameters using SHA3-256, then selects eligible provisioners. The important part is this: stake is not merely collateral sitting behind consensus. It becomes an input to the probability of being inside the decision-making committee. $DUSK reports 210 million+ DUSK staked, so distribution shapes participation.
That changes how I read “decentralization.” Many provisioners can still mask concentrated consensus influence when stake is unevenly distributed. Deterministic selection makes this relationship explicit and repeatable. The same inputs produce the same committee outcome across nodes, while stake determines whose capital matters more often.
There is a trade-off. Random committee selection reduces the need for every staker to participate in every round, supporting low latency today. But it also makes stake distribution a security parameter rather than a background statistic. If large balances cluster, committee exposure clusters too. Dusk’s audit history is relevant here: reviewers specifically examined consensus voting and slashing incentives, reinforcing that committee behavior and economic weight cannot really be separated.
I wasn't expecting the strongest security question to be “how many validators exist?” It may be “how evenly is security capital distributed?”
Does Dusk’s stake distribution remain healthy enough for deterministic sortition to preserve the decentralization its architecture assumes?

#dusk $ROBO $HEMI