Well, no hard numbers from NPD so far.
From what I understood, the ranking from the tool seems ok once again but the ratios might be off (I hope we will get a leak later so we can have a better idea).
I mean, if NS/PS4 real ratio is 0.8 (instead of 0.92 from the tool), I would consider it totally OK (expecting 100% accuracy would be ridiculous anyway). The only recurring problem the tool has for sure is underestimation of XBO. I will have to look into a solution to reduce this.
The real test will be July where the tool has NS winning by far.