Plan
Shortly, we are going to issue a chain variable which will increase the Proof of Coverage (PoC) interval from 30 blocks to 60 blocks. This will have the effect of reducing the rate of PoC requests sharply, which should decrease long blocks going forward.
Note that this may increase the number of low-reward epochs for some users, but it should average out to the same rate of rewards in the longer run.
Rationale
We initially picked the PoC interval of 30 blocks essentially because it was similar to our election interval of 30 blocks. It meant that most Hotspots would challenge each election period. However, validating PoC receipts remains challenging and computationally expensive. While this is something that we're working to ameliorate (see PR core/318), it doesn't really get to the core of the issue, which is that we are only beginning to understand the quality of the results from the PoC system. It's quite likely that even an interval of 60 is still far too often than is needed for good quality results, and since the PoC system represents (necessary) overhead on the functioning of the rest of the system, we hope to engage it as seldom as is possible while still getting good results. As such, we intend to spend more time in the near future attempting to understand system quality and to develop a metric that will allow us to understand how things like the challenge interval affect the quality of the results that the PoC system delivers. Our PoC system is quite novel, and we often make mistakes in its implementation. We appreciate your patience as we work towards making this system deliver high-quality results for all of our stakeholders.
Watch this space for more on the quality metric and potential future changes to the challenge interval in the future.