How do you calculate the average change? I would assume by
(Virtual Supply at the end of Q1/2024 - Virtual Supply at the beginning of Q1/2024) / days.

But isn't average growth still inflationary (albeit at a lower level)?

I should have been more clear on this point. It's almost certainly confusing to the reader. Sorry.

I jumped back & forth between two time frames in the text, but the charts in sections 5 & 6 show just one time frame. The visuals run from October 18 'til yesterday - roughly 5 ¹/₂ months, but I was only claiming that the blockchain was deflationary for the quarter from December 31 until I posted yesterday (and still today). On December 31, the virtual supply was about 502 million and today it's ~497 million, so we are deflationary for 1 quarter, but I didn't do the actual calculation to know the growth rate for that time frame. My only observation during the quarter is that the end is lower than the beginning, so therefore deflationary.

For the average new steem per day, I think I do the calculation per minute and then scale it up to days, but it's basically what you posted. Unfortunately, with the way I implemented the calculation in PowerBI, its not easy for me to change the start date, so I can't visualize Q1 alone. I have to give back all the dates from here. And you're correct. During that time frame, we were not (quite) deflationary. Which is why I wrote,

Even though the blockchain was deflationary from December until now, the STEEM price would need to move up a bit more to be deflationary over the whole time span that's delivered here.

To see the deflation visually during the quarter, the best place to look is probably the blue line in the top-right visual, here. Unfortunately, you can't see the exact dates, but you can tell that the virtual supply is lower now than it was in the December/January time frame. (or, you could check the raw numbers if you want to see the difference on the precise dates. Technically, skimming through the data now, it looks like we're deflationary back to about November 20 or earlier.) I had really hoped to get a visual of a negative average growth, but no such luck. It was negative during the entire period about a week ago, but it inched back up again before the quarter ended. 😞

I've been toying with the idea of setting up a testnet for a long time, where I can set various parameters independently of other witnesses and the real chain. Maybe I really should do that sometime to be able to play through various test scenarios.

This would be good for observing the technical mechanism, but even with a testnet, it would be hard or impossible to anticipate how interest rates would change stakeholder behavior, which is the real driver of price. I really wish @biophil were still here. I think we need economic/game theory analysis to have a chance at guessing. And, even then it might eventually come down to just running the experiment and being ready to make adjustments as we learn in real time. (Which is why I've been thinking about dynamically setting -- those values. But right now, I don't think we even know reasonable values for min/max settings.)