"As in the blue crop, ISO 200 is hard to distinguish from ISO 100. Which in practice is highly useful for that extra shutter speed. And it might well be that the Sony 8-bit compression makes ISO 200 truly as good as ISO 100 anyway (low level noise compressed away).
Most striking perhaps is the absence of any significant pattern or streaking noise even at ISO 25600 and even when the individual color channels are examined (including the red channel). In this regard the A7R *blows away* all Canon sensors of any resolution, and appears to be no less good than the Nikon D800E. In context, this is a true state of the art performance and speaks highly of sensor quality in the real world."
By the way...
If the DR of ISO 200 was smaller than at ISO 100 shouldn't the RAW Histogram in RAW Digger reflect that?
Without a clear description of Lloyd's methodology and the type of images he used for comparison purposes, I'm still none the wiser.
If both cameras at ISO 200 have similar SNR, tonal range and color sensitivity, as well as identical resolution, one would not expect to notice any significant difference in most images compared.
A difference in DR of 0.71 EV would only be noticeable in situations where ideally one needed to bracket exposures to create an HDR image, but was unable to do so for whatever reason, such as lack of a tripod and/or lack of time due to a changing scene.
In such circumstances a choice of ISO 200 might be preferred in order to get an adequate shutter speed for a hand-held shot, in conjunction with a good DoF at a small aperture. If one were using the A7r with an adapter, the circumstances where one would need to use ISO 200 would occur more frequently because one wouldn't have the benefit of image stabilization, even if the attached lens did have that feature.
Now it is certainly possible that DXO have either made a mistake during their testing, or they have accurately tested a faulty, or below-average camera body.
It would be interesting if someone could take the trouble to do a specific DR test at ISO 200, comparing shots of a very contrasty scene, such as the view out of a window on a bright day, exposing for the sky. The quality of the detail in the relatively poorly lit room would be ideal to reveal any DR difference of 0.71 EV, if such a difference exists.