I would have to mirror your exact thoughts and feelings put forth and an explanation after seemingly outlandish results from what by on the surface appears to be methodical testing would be nice. I would liken this to producing what appears to be a great widget, but having no support for that widget once it leaves the factory. I see both sides, as someone who's worked online for yours, it would be impossible to address everything every time a subscriber finds issue, but some times it just becomes imperative that you do. Some one like myself and it appears, as yourself, will find fault and without additional explanation, one has to disregard all observations and data put forth. As such, I will no longer take the time to read or travel to the website, which is how the site pays the bills, so to speak.
In the last six months, I've purchased three expensive laptops, the least expensive being the Google Pixelbook. I work online and use my laptops a minimum of 4 hours a day every single day of the year and it's usually more like eight to twelve hours a day on a normal "work" day. I know without a doubt that my Pixelbook can get six to eight hours of mileage out of a full charge (brightness at 100%). I not only use Chrome a lot, but I do significant amounts of photo editing with the Pixelbook. If I use Chrome a lot, accessing various web apps, I get closer to six hours than the closer to eight hours I get when just using apps. Even when dialing down brightness to 50% (which is the minimum I can go for most use cases), I can stretch those times one to two hours. I've repeated results multiple days over and over since the laptop was released late 2017. If I had based my purchase on this article, I would minimally be disappointed and most likely angry I spent $1000 USD and getting not even getting 50% of what is claimed here. As well, I am lucky I did not base my purchase on specific assertions and claims made in reviews of the Dell XPS 15 9560 and others. As mentioned previously, it's not the entire reviews, it's specific individual results that seemingly could only be different hardware or potential issue with methodology in testing that I to feel need addressed and followed up by NCR.