They apply arbitrary - and by their own admission, completely subjective - scores to a series of metrics to end up with one number.
The issue is that if you don't give the same priority to the metrics that they do, the scores can be moved substantially.
The simple fact is that DxOMark's conclusions are no more objective than simply looking at the images and picking the one you like best.
That's true of the aggregate use case scores. But we are discussing the graphs which do indicate unambiguously that the D800 sensor has better dynamic range at low ISO. Whether or not dynamic range at low ISO is important is subjective and depends on how you're using the camera. How to weight the three different use case scores to get the final number is also very subjective (to some users only one of those three measures might matter). Also, they made some somewhat arbitrary choices about the way they combined the different factors to get the high ISO score.
To get a more accurate picture, you really need to look at the measurements to see where the differences really lie, and this is what was done in this thread.
We won't know exactly how good the 5DIII is until we see the scores for that, but it's pretty clear that the D800 has a better performing sensor than the 5D2 (but hardly surprising -- previous Nikon models were already performing better. Even the Sony APS-C cameras were doing better for low ISO dynamic range)