In that article, both cpus at stock, the pentium won 3 tests and the Athlon 4, based on frame time variance, although when the Athlon won it did tend to be by a bigger margin. So it wasnt a definitive victory for either cpu. In any case that was with a discrete gpu so is not really relevant to this discussion.
I didnt say that it was a flawed conclusion from THG but that they exagerated the pentium strength in the first review, to be declared a honor product generaly mandate a clear cut in the scores wich, as you point it, wasnt the case, also you re right about the dGC rendering this exemple pointless for comparisons in this thread.
