My thinking is this was always going to be a challenge. My question above is more to clarify that there was a belief that if Scorpio supported packed math it would have been able to show obvious differences, but now that we know it doesn't have that support, the belief is that it won't?
I never thought it would be as obvious as they claimed, but there was a good argument that even if it's somewhat a small difference, like between ps4 and xb1 with a lot of 900p vs 1080p (so maybe 1800p vs 2160p, or CB vs native, whatever it would end up with), the media hype and side by side comparison would have an impact to convince it costs a little more for a little more IQ. It was already a borderline proposition. For $50 more it would be well accepted as great value. Which was my guess.
Now it looks like the 45% advantage will be smaller, let's say it becomes 25% (made up number, I am guessing wildly), it's like 1800p versus 1900, on - consoles - which is already a market where not many gamers care about absolute performance.
If they take a small loss (like 50? 60?) on the hardware to get it down at the same price, they gain market share by offering more for the same price. There is no more comparsions where gamers ask "is this really worth $X more?" it becomes a no brainer advantage, a bullet point in a list of advantages without caveats. I think it's what caused the Pro to have a pretty average launch and gamers still prefer the lower cost slim. It's much more powerful and still wasn't worth $100 more.
I am only seeing this as an angle for MS to regain market share, regardless of how many upgrade to scorpio. Of course it will sell well to upgraders regardless, they are getting a bigger jump so it could be more worth the price than it did for ps4->pro.