Might I remind you that better = better, does it not?
In this case, no it does not. I'm using the performance of the crop sensor to gauge the performance of Canon's on-chip ADC, not the performance of the sensor as a whole.
In this case, better = better per unit area of sensor, not per whole sensor. Otherwise you could just say that medium format trumps everything, by virtue of being larger (practicality and lens availability aside), despite CCD sensors being, area-for-area, worse than almost any CMOS sensor currently on the market.
I disagree.
I am pretty sure Nikon leadership is smarter than you.
You haven't the slightest clue who I am, so don't presume anything.
As I showed on another thread the D810 owns the low ISO area (or did at its release), and Nikon is therefore letting the D810/D900 handle the "low ISO fans" ... while creating the D5 to own the high-action/high-ISO department ...
You haven't shown anything except empty rhetoric. Where is the data that backs up your claims?
So far, none of the graphs you've posted actually back up anything you've said - all of them have basically compared the D5 to old Canon cameras, then somehow drawn the conclusion that, since they're better than the old cameras, they're also better than the new one.
Nikon hasn't "gone backwards," what they've done is create two different, specialized tools, each leading the pack in their respective areas, as this graph shows 
Then explain why the D4s has measurably better low-ISO performance than the D5, while only losing by half a stop at high ISO. You'd expect at least a half-stop improvement in several years of sensor evolution anyway, so it's hardly that they've sacrificed ISO 100-1600 just to gain half a stop at high ISO. That's going backwards by any measure.
Like every other company, Nikon is trying to maximise its profits, not create the best sensor or the best camera. This means making optimal use of its resources.
Nikon does not make its own sensors - it designs them and gets someone else to make them. On-sensor ADCs add an extra layer of complexity to the sensor. More complex sensors require more advanced fab plants, which sell their services for a higher price and so cost Nikon more money. These sensors are necessary for a non-action camera where detail is everything. For a low-resolution action camera, which primarily lives at ISO 800 and up, it just makes for a much-more-expensive sensor without greatly benefiting its primary purpose. So, skimp on the unnecessary parts of the sensor and pocket the change.
Nikon owns the base ISO and the high ISO territory ... while Canon is "champion of the middle" 
Your graph doesn't show anything about the 1Dx2, nor can anything about the 1Dx2 be inferred from it. All it shows is that a brand-new Nikon sensor beats the 1Dx and 5Ds at high ISO (being about the same at low ISO) and the D810 beats both at low ISO.
We are looking at the performance of the 1Dx2 as compared to the D5, not the 1Dx and not the 5Ds. The 1Dx2 sensor is not the 1Dx sensor, and not the 5Ds sensor.
All we know for sure about the sensor at the moment is that:
- It is a 20MP sensor
- It uses on-sensor A/D conversion
Also, from being 4 years newer, it would be reasonable to assume that high-ISO performance has been improved significantly - say, by about a stop - as this is a general trend that occurs with every camera.
It is demonstrated, from the 70D/80D example, that on-sensor ADC linearises the low-ISO SNRs to match the relationship seen in the higher ISOs. If that were applied to the 1Dx sensor,
even with no other changes, it would already give a DR of 11-12 stops (by the 20:1 SNR threshold). So, just that one change, without any other upgrade to the four-year-old sensor, would have the sensor beating the D5 below ISO 800 and matching it up to 1600. Then it only needs half a stop of high-ISO improvement - very reasonable over 4 years - to beat the D5 at high ISO as well.
There's almost no way the D5 can have a better sensor than the 1Dx2, assuming the measurements on the website are correct and what Canon has told us about the new sensor is also correct.
"We?" Do you have a mouse in your pocket?
You are projecting some "new" leadership in Canon, who (in fact) now has only 1 camera in the top 10 of two classes (the 5DSr), which means Canon only occupies 1 spot in 20 available spaces, and it occupies the bottom rung of FF DSLRs at the moment ... while not a single Canon shows on the top 10 of APS-Cs ...
Nikon, meanwhile, owns 10 of 20 possible "top spots" in DSLR camera positions ...
I'm not projecting anything onto Canon. I'm merely pointing out what Canon has already demonstrated in the recent 80D - a huge improvement in low-ISO SNR. The numbers are there in front of you, for everyone to see. And the new 1Dx2 sensor uses the same technology, so one would expect the same jump in performance.
Past statistics are meaningless when one side introduces technology that significantly alters their cameras. They hold some weight when you're talking about slow, evolutionary change, but the shift to on-chip A/D conversion is not evolutionary change.
Just look back a few years and see how badly Nikon sensors were trailing Canon. Canon had the 1Ds2, 1Ds3 and 5D2, while Nikon couldn't even produce a full-frame sensor, or a half-decent CMOS. If you made a list back then, pretty much all the top sensors were Canon. Then Nikon got their hands on Exmor and everything changed.
My thinking is Canon will fall short again, based on the overwhelming evidence.
The plural of 'anecdote' is not 'evidence'.
Again, I think you are failing to realize what Nikon is doing: leaving base ISO scores to the D810 (and, soon, D900), while making the D5 the tool of choice in the high ISO world ...
They're not doing that at all. They're segmenting cameras by application, not by ISO capability. That would be a pretty dumb way to segment cameras, since action photographers still shoot ISO 100-400 in good light, and non-action photographers still shoot high ISO when shooting non-action in dark conditions.
You are stuck in the insanity of believing "one" camera will ever be "the best at everything" ... which isn't going to happen.
I never said that. Quit putting words into my mouth so you can argue against straw men.
Nikon is committed to producing the best base ISO camera, which it did, and producing the best high ISO camera, which it did, leaving the rest to be mediocre.
The D500 will be the best APS-C, the D5 the best high-ISO/action FF, and the D900 will once again be the best base ISO FF (that is my prediction).
Are you a Nikon executive?
If not, how do you know what Nikon is committed to producing? All that can be assumed is that Nikon is committed to maximising its profit.
Even if they were committed to producing the best of everything doesn't mean that they are actually going to achieve it. There are other companies out there who are equally keen on beating Nikon, and have many times the resources (Sony and Canon).
No. You just need to wake up and smell the coffee ... and finally recognize what's really happening 
Either you have no grasp on logic, or no grasp on mathematics.
You said, "Sure, the D4s was better than the 1Dx - but it's also a year-and-a-half newer. D4 and 1Dx were more-or-less comparable," which was essentially parroting what I said about the new AF systems: Nikon's will be better again (and it is), though only marginally.
What do you think 'D4 and 1Dx were more-or-less comparable' means?
IT MEANS THAT THEY ARE EQUAL, not that the D4 is marginally better than the 1Dx.
The D4s doesn't count. Canon didn't bother releasing a competitor against it.
In other words:
D4 = 1Dx
D4s = no Canon equivalent
D5 = 1Dx2
On what basis are you saying Nikon's will be better, anyway? There's no track record of one company's action camera having better AF than the other's, when both were released around the same time. D3 = 1D3, D3s = 1D4, D4 = 1Dx, so it's likely that D5 = 1Dx2.
We agree on the video disparity favoring Canon ... only.
On the rest, we disagree. Completely.
Do carry on though 
Jack
You can't argue with fanboys.
For the record, I primarily shoot Sony.
