These are really good points, @s_ben. It aligns with some advice @jbenet gave me around the importance of decoupling the rewards a bit.
For the CredSperiment, I was originally planning something like 50/50 (last-week/all-time) payout ratio. But I’ve wondered lately if it would be better to keep a smaller proportion on the immediate payout, and let more of it go through the adaptive and long-term system. Your comment inclines me more to go with the slower system, maybe 20 immediate / 80 long term. It’s more conservative, and we can always go more aggressive/faster later on, as we get more confidence.