Well, the games I usually play are almost never going to be included in any review, so I am forced to use other games as benchmarks, which in itself is not a bad thing. Yes, it's not easy, but neither is it impossible to have an apples to apples comparison between video cards.
The problem with B3D reviews is that while I'm sure they are very technically advanced and accurate in their assessments, for the neophyte 3D enthusiast like me, half of the review usually doesn't make much sense. That is why I have to look mainly at the benchmarks and the conclusion. And to make any sense of the benchmark, I have to compare the numbers to the benchmarks of the competing product. A 65fps in UT2003 doesn't tell me anything unless I can compare it other products, either in the ATI lineup or even Nvidia's' offerings.
Right now, I am forced to go to other website to get such comparative results. Now, I do realize that B3D caters to a niche audience who better appreciate such a review format, but I really hate having to go to hardocp and the like to get my reviewing numbers. Who knows how accurate there results are.
Here's an example of a shootout I have in mind. It gives me a good idea of where ATI and Nvidia is in relation to each other, but again, without the authority that B3D's work carries.
http://firingsquad.gamers.com/hardware/ut2003perf/default.asp