There is a single data point (blue cross), and it tells you that the quality curve of the encoder is probably going to be significantly better than VP9.
The author correctly called it: "a single (admittedly non-useful) data point."
it is consistent with previous results
I see no "previous results"...
That basically means the situation will repeat as with H.264 and VP8.
That's nonsense. VP8 was only released to the public MANY, MANY YEARS after H.264. Now VP9 is there right from the start, on equal-footing with H.265. This is a very different situation that EVER before, which was the point of my original post you apparently dismissed out of hand.
n the MPEG side, the competition will cause the actual practical quality of encoding to use much more of the theoretical potential of the format; whereas on the VP* side, there will be just one not very progressing encoder library,
This is utter nonsense. The development of MPEG encoders always follows the same path... There's innumerable proprietary ones, with absolute CRAP quality, and then ONE open source codec which gets all the development effort and blows the rest away. With MPEG-1/2 this would have been mjpegtools, with MPEG-4 ASP this was Xvid, with AVC this was x264, and with HEVC this looks to be x265, though that could change.
The fact that there's a single open source VP9 standard means there will almost certainly be less duplication of effort, and the single central open source standard will get all the high-quality coding improvements.
And you're basing your whole opinion on this completely flawed logic that flies in the face of reality?