The thing is - I'm not defending the position of whether Civ7 is doing fine or not. I promote healthy skepticism about the data we have. Basing any significant conclusions on a metric which was designed to measure multiplayer health, is not great.
Regarding the simultaneous player number metric. We know for sure is that Civ6 did really well post launch, reaching 1M copies very fast, but this fact has zero reflection on the graph. It has the same drop curve as Civ7 which presumable had bad post-launch sales.
I understand the desire to use it because it's the only freely available metric we have, but it doesn't have significant correlation with any commercially valuable metrics like sales (and that's why it's freely available, BTW).
Well, you pointed out that Civ6 peaked on second day instead of the first and that's the primary explanation.
We know that it did worse by Civ7. Because Civ7 did better.