The biggest problem is that it's anecdotal. From my read, the "14 hours" was the reviewers daily use of the phone, which is useful, but only marginally so from a comparison standpoint.
Sites like Engadget are using a battery drain script (web pages reloading or video or something), so even though they are not transparent when it comes to the methodology, at least we should be able to theoretically compare common performance across phones.
What we really need is transparent methodology (why has no review site released an app that allows users to test battery and develop crowdsourced battery life information?) and in addition multiple tests of the different variations of radios on/off (wifi, BT, LTE, 3G) across multiple phones to really get a sense of the battery impact.
...which means the closest we'll get is having to wait for the AnandTech review whenever that one comes out.
System test is an engineering discipline. Because it's only touched on in most programs, many engineers leave school with limited knowledge, if any, of the discipline and the art, and take years of on the job training before really getting it. Lack of the discipline is why so many bugs make it to the end users.
As for a crowdsourced app, I was going to post bravo, then I was going to pick at the idea, but instead, I am intrigued and will say this - we would need a way to discard the results from those who use their phones wrong, task killing and social networking with redundant polling, and also those trying to game the system in their favorite model's favor.
But, it occurs that with a sufficient population to sample from, those two extremes might tend to cancel each other out.
So, here's what I think - Phandroid and AndroidForums.com sponsors the largest crowdsouce app for carrier coverage. (Some hate us for it and say we're in the carriers' back pockets, but that's not true.) I am going to suggest that we launch a similar app to crowdsouce battery life. No promises, no guarantees, and everyone agrees in advance to not be upset if we can't swing it, ok?
Meanwhile, I do like these anecdotal reviews. They never match my use or lifetime but I do get a notional view of things (of those that I don't reject for outright silly methodologies).
Meanwhile, I agree, can't wait for a full AnandTech wrap-up on this pup, especially for battery life.