When a healthy verifier is deactivated but maintains sentinel protection, that verifier will ride around the cycle for several days before reaching a score that justifies removal. When a person or group that controls many verifiers deactivates all of them at the same time but maintains sentinel protection, the cycle can have throughput problems for some time, and the frozen edge can fall behind the open edge. This is exacerbated by protections that were implemented to avoid removing verifiers from the cycle too quickly.
This is how the system was designed. The effort required to get into the cycle is respected. Hastily removing verifiers from the cycle would have its own drawbacks, making it easier for attackers to force verifiers out of the cycle and making maintenance of verifiers more demanding.
The cycle is good at catching up, though. It often achieves block verification intervals of less than 3.5 seconds when behind the open edge, and almost all verifier-produced (non-sentinel) blocks have verification intervals of less than 5 seconds in this situation. This allows the cycle to erase a deficit of about 5,000 blocks per day when behind.
The deactivated verifiers will eventually be removed, and the cycle will catch up. Nyzo will continue stronger than ever before, and new verifiers will take the place of the departed verifiers. This is how Nyzo was designed, and this is how Nyzo has proven itself to work in practice.
Departures are unfortunate. We do not like to see people leaving the project, and we wish they would see the potential they are leaving behind. But we wish them well, and we thank them for their contributions.