@Thogard thanks for your thoughtful analysis. I recall reading some of your ideas about re-orgs previously and they were equally helpful.
Regarding your questions about the sprint length change, there is no immediate impact on checkpoints or explicit finality. Obviously, if the rate and depth of re-orgs decreases as intended, there is a significant benefit to probabilistic finality and this is part of the motivation for the change.
Having said that, we are also planning a subsequent set of changes - which will require a hard-fork as well - that will directly address this and formally anchor finality to sprints. The intent with this is to have much faster, more explicit finality. More on that soon.
With the change to the base fee calculation, I would say that spam transactions are part of the issue but only part of the story. While spam is often a problem some of the spikes we have seen in the base fee recently have been driven by sustained surges in usage from legitimate applications.
I think we can all agree that the robustness and success of applications is the primary driver of value to any blockchain - all the more with Polygon given the range and diversity of the applications we host. I’d wager if you polled any number of the prominent application teams on Polygon they would say the wild variability of gas fees is their number one current challenge, even above re-orgs. We’re aware of some teams that have had to implement extraordinary - and unhealthy - mitigations to manage the spikes. From our perspective, some of these spikes are tantamount to outages - the system functionally becomes unusable.
Moreover, our modeling and testing has shown us that the core EIP-1559 parameter - which arbitrates the maximum rate of change - is just too high for us given the much greater throughput we have compared to other systems like Ethereum. If anything, I’m concerned that the 2X change we’re proposing here might not be enough - but we’re intentionally being cautious about this.
Finally, I take your comment about legacy gas pricing as being somewhat informal. 
Any gas auction mechanism is tricky and needs to balanced correctly. Basically it’s a hard problem. But EIP-1559 has proven to be successful at managing capacity on Ethereum while improving gas price stability and I think it can be successful on Polygon. We also consider it critical to maintain compatibility with much of the tooling and infrastructure in the ecosystem. We just need to get the tuning right.
Thanks again.